SAY NO TO RECOMMENDATIONS

I have gone tired of telling class after class of public servants and professionals to stop using the word recommendations at the end of analysis. Recommendations are lazy, vague and distant attempts at solutions without any responsibility and without any pathways. Recommendations are inherently preachy, have no timeline and no work distribution. If analysis concludes with recommendations, you can assume hundred percent that it will not be picked again. If on the other hand, the reader (or policy maker) finds action items, tangible steps and clear plans, he/she will stop and pay attention.

I have had a chance to engage with private sector and have seen how they are trained to solve problems. Besides the point that they always appear paranoid and almost freaked out, the talk is always about the specific steps and specific interventions. They don’t take kindly to recommendations, and I can’t imagine a senior executive surviving on job after presenting a wish-list of recommendations. In public sector, such executives thrive.

INTERVENTIONS or PLANS

I am more in favor of the term intervention as solution because the whole framework rests on the assumption that the system is going on its own pace, has a gap which needs to be filled, and something needs to be done to re-direct towards the problem. Changes are needed for the course correction. These changes mean intervening into the systems and alter the way the things are being done as per now. It’s time to break the status quo or inertia and introduce some interventions.

Plan is other option. What constitutes a good plan, I always ask the participants in training workshops and what I get is answers in the range of SMART Specific, Measurable, Achievable, Relevant & Time bound. The thing with SMART is that these are attributes of a goal, plan is to reach these goals, and that plan will not make itself. Someone has to come up with a plan.

I think a good plan has these seven qualities

1. Clarity - in both roles and goals. Who will do what and what are the targets - across time and domains

2. Risk management- things are not expected to work as per plan and there will always be risk. A good plan pre-empts the risk and mitigation strategy.

3. Team- that will implement the plan. In public sector teams are disparate and spread and only comes for meetings which leads to lack of ownership.

4. Ownership- whose head will be down the line and who will share the accolades

5. Monitoring- Not the usual M & E but tracking of progress and quality by someone who is not implementing the plan but has ownership

6. Resource mobilisation- plan needs new resources which needs to be mobilized. Not just budget and finances but the legal, human and social capital.

7. Parameters of success- Plan is not just about inputs and activities. How is success defined and how is it measured, is essential for a good plan.

WHERE WOULD THE SOLUTIONS COME FROM?

Most of deliberations (meetings, consultations & reports) in public sector are exercises in selection of a good solution, like selecting dishes from menu. Policymakers also like to be presented with a menu card of solutions from which they choose as per mood. These solutions are oft repeated, convenient and reactive measures which are presented and adopted because the policymakers need to ‘do something’ to address the problem.

The solutions should logically and ideally come from the problem. Our problem analysis invariably has acquired the shape of gap analysis. We have a problem point (Point-A) and Solution point (Point-B) and worst-case scenario point (Point-C). BY now we have mapped the difference or gap between the two points A & B. Our interventions should focus on the actions on two fronts: removal of problem drivers and provision of catalysts.

• Removal of problem drivers: Our problem is persisting due to certain drivers and sources. It can be as simple as lack of monitoring and as complex as changing the culture, depending upon the problem

• Provision of catalysts for change: Change does happen by removing of the undesirable. It has to be sustained by capitalizing on encouraging forces.

The idea has been taken from the Force field analysis in change theory. Force field model is a simple but powerful framework assumes that there will always be two sets of forces: those helping and supporting the change versus those hindering and stopping it. The plan should focus on capitalizing the encouraging ones (catalysts) and removing those driving the problems(obstacles). Any intervention which does not do that will be meaningless and probably waste of resources.

DIVERGENCE-CONVERGENCE MODEL

I call this model overall divergence convergence model. We started from one problem point (point-A), we diverged in our exploration phase. We then focused on the situation and the context and moved backwards and identified the problem sources. These problem sources have led us to the options. But we cannot go for all the options because a. it will be too cluttered and b. it will not represent the reality of problem. Even in creating options we can diverge even more beyond four or five easy options. This phase is the creative phase. This is the time for some out of box thinking and come up with combination of options which will solve the problem.

Then comes the time to convergence because the analysis has to be presented to the client be it the Minister, CEO or even the general public-they want a concrete proposal. That requires zooming in, which can only be achieved through the selection of neat and concrete solution which encapsulates the problem and is reasonable, relevant and realistic or 3 Rs.

Let’s get back to the causation exercise we did in the previous sections and extend it to create options. Each option will arise for the particular problem source. One source might have maximum contribution in the perpetuation of problem, others not so. The analyst through his/her insights, available data and logical framework would have an idea about the relative weight of problem sources. The combination of choices will lead to the solution or Point-B

HOW TO CHOOSE THE BEST INTERVENTION

The acronyms are always in fashion and for a better (one can never be sure of perfect) solution should meet the criterion of three Rs- Reasonable, Relevant and Realistic

Reasonable; the intervention should have link to the drivers of problem, the source and the elements which has sustained the problem, in nutshell is linked to the problem.

Relevant; Relevancy is drawn from the context and situation. Is the intervention a case of applying apples to oranges, or is it like a round peg in round hole?

Realistic; A realistic intervention is not just economical and efficient but also low risk. It has better chance to adopted and scaled by the recipients.

IMPLEMENTATION PHASE: THEORY OF CHANGE

We have come up with a solution, now is the time to explain how it will be implemented. Theory of change is popular concept in the development sector. Not that it is panacea or magic bullet which will keep the change programs rolling but it is useful in keeping the plan on track.

Theory of Change is essentially a comprehensive description and illustration of how and why a desired change is expected to happen in a particular “context”. It is focused in particular on mapping out or “filling in” what has been described as the “missing middle” between what a program or change initiative does (its activities or “interventions”) and how these lead to desired goals being achieved.

In the simple table form, the TOC will look something like this

One can add INDICATORS on the horizontal axis to measure the progress across inputs, outputs, outcome and impact. I have added my own contribution to the Theory of Change, which is the future reality. Hard to quantify but the analyst should have a vision for future reality. Noy dropped from sky but developed through the process of analysis.

Future is all about uncertainty which is driven by the change in situations and contexts and leads to unpredictability. Analysts are in business because of the uncertainty. The source of uncertainty is always incomplete and imperfect information. If we have perfect information, everything will be certain. Due to uncertainty, we cannot have exact trends and patterns analyses for the future. At best we can extend the trends into future but that will be based on the fundamental premise that past will be exactly like future. It might actually be, as is the case in number of instances at personal and society level, but can one be sure about that?

Uncertainty and limited information are the realities of life. I recall Henry Kissinger’s statement somewhere (though not sure where exactly but remember the idea) that policy maker will always have limited information and will always be faced with uncertainty. If there were no uncertainty and perfect information, the decision would be obvious and clear. Kissinger, the superior mind he was, is making two important points, a. there will never be complete information and b. analysis, and consequent decisions will always be carried out in uncertainty.

To predict future has now become an industry. Books after books are coming out on the idea of the art and science of forecasting. Tetlock and Gardner’s Superforecasters; The Art and Science of Prediction is good primer on the imperfect science. Then once the darling of political forecasters but lately out of color Nate Silver’s Signal and Noise; Why So many predictions fail and some don't was promised to be the last word on forecasting. Not surprisingly the promise could not hold. Predicting future was never supposed to be easy. Some experts like Nasim Taleb cautioned about the improbable with works like Black Swan : The Impact of the highly Improbable

We are here more focused in including the future into the written analysis before we reach the holy-grail of policy- How to change the future. We should also keep our scheme intact- first the problem definition then situation & contextual analysis followed by the causation and then taking the full circle to future and finally the interventions of change.

Let’s get back to some of the problems that regularly require predictions.

• Economic indicators like GDP growth, inflation & unemployment

• Political questions, who will win elections, who will survive and who will fail

• Social indicators; education, health and climate

• Security and violence related; will terrorism and extremism surge or taper off

• Business and enterprise; how will AI affect companies, what are the technologies of future and who will be winners?

If we look at these issues, we will see that some are more predictable than others. Economic indicators are normally more predictable than the political ones. We can make better predictions about a country’s economy than its politics. Social indicators at societal level are also within range of prediction. One can make a guess regarding India’s social indicators in comparison with its neighbors like Pakistan and Bangladesh. Business and companies used to be predictable before the information technology but not anymore.

We can also take a look at the recent developments in the economic, business, social, security and political fields and see how accurate the predictions have been. Let’s take three monumental events that have shaped the world in last year or so- Middle east conflict, Trump elections and emergence of AI. Despite so much predictive knowledge capital in the world, the actual range of effects could not be grasped.

• AI has been in news but then came OpenAI and ChatGPT which blew away the businesses and sector like never seen. Even when the world was grappling with GPT-4, came the Chinese Deep seek. If the pundits are expected to take a wager today on what is future of AI in next 5 years, the range would be wide by margins.

• Donald Trump was butt of jokes at White House at the hands of President Obama in 2011 and in 2016 was elected to the White House, got defeated in 2020, started a near insurrection and elected again in 2024. Even when his second term started looking like a reality, no one could predict what it will look like. EU and NATO despite so much talent at their disposal could not have in their wildest imagination have guessed that Trump will pull the plug in matter of weeks. All those US-Aid employees who were making plans for their future were without jobs in 24 hours, and many more in the government too at the hands of DOGE.

• Covid-19 pandemic before that was predictions failure galore. It came so sudden; it impacted even more suddenly but the impact was uneven. Could anyone have guessed that more people will die in London than Lahore? When will be the next global pandemic come, and how will look like, take a guess.

Prediction specialists have a convenient explanation, black swan events. Indeed, some events are like black swans, if you haven’t seen one, you cannot imagine even. But the real failure lies in adjusting to the reality and forecast even when the black swan has emerged and been seen. More than often, events take their own shape and not always for the worse, and the forecaster keep on swinging between doomsday and comforting.

The predictors in the field fall in two categories: Doomsdayers & Comforters

I have avoided using more familiar terms like optimists and pessimists because these terms represent more the state of mind about a situation in present. Also, optimism and pessimism are more personal oriented. What we are looking for is the capacity to assess system impersonally. Predictive capacity is more like a habit, a muscle and a reflex action. One can be an optimist in prison and a pessimist in the luxury hotels but there will be certain people who are optimist in general outlook and still paint a bleak future for others. The distinction is between those who think that things will go worse and those who expect them to continue as they are. I call them doomsdayers and comforters.

Doomsdayers are always preparing for future worse than present. They are different than pessimists because they don’t resign in the face of expected adversity but are actively working to avoid it. It’s just that their version of future is worse and without any basis. Not that the future can’t get worse but if you want to ask them the indicators on the basis of which they have reached doomsday conclusion, they will struggle to find cogent answers.

Doomsdayers are born risk averse, have seen personal ups and downs and are shaped by societies that have seen upheaval. They always have a story to justify their understanding of future even if the context is totally different. Some actually get motivation from doomsday painting of future because they think it keeps them on their toes. Doomsdayers in a way are followers of saying ‘prepare for the worst and hope for the best’, though the former has more weight than the latter.

We have fair share of doomsdayers. Most of our elite falls in doomsdayer category. One hears in the country’s capital regularly about how the country is failing and why this is the time to move. They also share their plans of taking their assets and shifting their families. Then they have number of examples of notable people who have already shifted. Number of cognitive biases are at play here, anchoring, bandwagon, slippery slope etc.

Comforters are at the opposite end of spectrum. I have given the name comforters for their desire and hope for comfort of the present. Comforters have also seen a lot and have the realization that monumental events which affect the societies take long time to come to fruition. In their favorable expectations, they also miss or minimize some clear signs of the problem. Their solutions also reflect their thinking. Comforters are status-quo oriented and do not believe in disruption. Not that they are fiery optimists, but more inclined to take and expect things slow.

In the case of out of school children, the doomsdayer will draw a future in which millions of children turn into criminals and extremists. On the other hand, comforters will rationalize the present with argument that out of school children are actually earning for their families as compared to the situation where all are educated and unemployed.

Both doomsdayers and comforters lack the basic requirement for prediction pathway to the future state of affairs. What will be the road from Point-A to a future point (lest call it point-C) look like. How much time will it take, what events will happen and how will people with agency will respond to events. Yes, the countries go through crises, some countries do fail and collapse, but the forecasting analysis need to have the pathway from point of failure. Not just what failure will look like, but the road to the failure.

Who are the realists?

Talking about the Doomsdayers vs Comforters, Optimist’s vs Pessimists and Ostriches vs Over-prepared, one is always reminded of another specie the Realist. Realist is an imaginary creature who cuts through the noise, locates the facts and makes decision on evidence. Realist is also portrayed as calculated, emotion-less and lacks empathy and blamed to be only watching for their interests without regard for the human misery. Kissinger was known and blamed to be one cold blooded realist.

I think every one of us want to be realist from inside and passionate from outside. We want to make decision on facts and evidence as much as possible but want to sell on the basis of emotions and passion. We know that reality matters more but we also know that people are moved by passion more than reality. After all people are also doomsdayers and comforters. The clever ones play with the two strands simultaneously: creating doomsday scenario to create fear and exploiting the comforting side to create hope, both according to the situation. You might have seen many popualist leaders doing that.

Back to the key question ; Where will we go from here?

We should be mindful that we are not starting from nothing as most analysts do. We have a well-defined problem, we have mapped the situation and context and also have an idea how the present emerged from the past.

Lets use the graph we have been using in the previous sections.

The addition here is the Point-C. Both point B & C are in future but point C is the expected future if we don’t do anything while point-B is the desired future after our intervention. Point-B is thus our goal or target which we intend to achieve through efforts. Point-C is our fear and point-B is our hope.

How do we know our Point-C.

Future is uncertain and we have to start with the premise that there can be more than one future. Such multiple futures are driven by multiple scenarios. The common and convenient approach is thus the scenario building. Scenario-building in simple form will look something like this

Multiple futures can be mapped in multiple scenarios. These scenarios emerge from an idea of the range for the future. Think about the problems and issues we have discussed earlier.

• A developing country can Pakistan’s range of growth rates is between 2-4 %

• The company’s sale range can’t go beyond the production capacity and the total market and expected range is between the market share and capacity.

• Out of school children’s range is determined by the population rate and average enrolment ratio.

• Inflation range will be between the money supply and supply of essential commodities.

If we take a close look, the range is determined by two factors, default limits and growth potential. Economic growth has the default limit determined by the factors of production. A company can at best be the monopoly, the only producer but it cannot go beyond its supply capacity and available demand. Inflation has the widest limits imaginable, and countries have seen hyperinflation when the currency can lose value to the millionth degree in short time.

To map future is essentially to also measure the growth potential. Out point A is on a spectrum and has been long time in the making. One is always tempted to assume that the trend will continue. if a country is having 2 percent growth, then one would be cautious to assume it will be 3 percent. If a company has market share of 25 percent, one will assume that it will be around 30 %.

In all fairness, these growth-based estimates are reasonable because these are determined by the default factors which always have a limited range. Thats why planning is also optimistically adjusted, be it in the case of country or a company. But there are two factors which makes this whole exercise redundant: volatility & Time Horizon.

HOW TO THINK PROBABILISTICALLY

You might have used the words like likely, highly likely and unlikely, mostly without much thought. We have also used expressions like MAY, MIGHT, WILL, SHALL etc. Again, our anchor is what has happened in the past. If it has happened before repeatedly, we will call it highly likely and will be confident in saying that it WILL happen again. On the other hand, if something has not happened, it will be unlikely, and MAY or MAY NOT happen again.

The question of prediction is to identify the room for a TREND or NEWNESS. Lets take the example of movies. One action movie is crazy hit. Producers think that action movies are now a trend and starts a big budget action thriller. It is tanked at the box office because the fans already have had enough of action. They actually craved for a romantic comedy or Rom-Com, which turned out to be a low budget which no one had paid attention too. Same thing applies to politics, security, business and society. Do people at particular time and space, want more of the same, or are tired of the same old, really depends on the Analyst ability to think probabilistically.

My proposed framework is POSSIBLE-PROBABLE model. First identify and map what is possible. The four scenarios mapped above are four possible outcomes visible from point-A. Other than these four, one should be confident that everything else is impossible. Once you have ruled out the impossible, what is left as possible is probable. But how probable?

If I were confident in my understanding of Bayesian Framework or Bayesian Epistemology, but I will restrict myself to think in terms of odds for now. To understand odds is to exclude the impossible find the most probable. Let’s take examples we have been engaging with, out of school children. What are the odds that an educated person with monthly income of more than one million, will not send children to school? What are the odds that an uneducated person with monthly income of thirty thousand will send the kids to school?

On both accounts, we can say with confidence that both the cases will be in the category of highly unlikely. Both are good samples of behavior of their respective groups. But there are always the outliers. I keep on reflecting on the arrival of passengers at the airport for international flights departure. There will always be some passengers who will arrive four hours early (confession: I am one of them) and then there are those who will come just an hour before the time juggling traffic and battling the queues. But most of passengers will come 2-3 hours window before the flight. If you plot these arrivals on a graph, it will start looking like this

The title says it all normal distribution or the bell-shaped curve is the most common, most popular and indeed most universal visualization of any data from any sets of population. In some cases, the curve will be flattened meaning that number of scarious are more and evenly distributed, in other cases the bell will rise vertically showing more variation between middle and ends.

To think probabilistically in case of lare popylation data like eductaion, health, sales and voting, is to think about the future normal distribution. Most of the analysis are about the collection/convergence of large data point as opposed to one person’s actions. Though latter matter when that person is incharge like the US President Trump, or CEO of major company, or head of the terrorist organisation. Even those are shaped by past patterns and future needs.

Sports competitions are common and relevant examples of probability thinking. In the soccer match between Argentina and Cameroon, any football match knows the odds but also know that Cameroon odds cannot be zero. In the past Cameroon has beaten Argentina, the team is well prepared and Argentina on the other had string of injuries. An objective and realistic analyst will put the odds at 80-20. 70 % is more likely and probably the Argentina will win comfortably. It doesn’t mean that odds are wrong but that if both teams play 100 matches, Cameroon will win in the range of 15-20.

Thinking probabilistically is hard because mind works in binaries 0-1, a thing will happen or not happen, and if one thing happens, the other does not. Yes, at the end, it is always one thing that happens- someone wins and someone loses, but not that loss was 100 % destined otherwise there were no game. Players go into game with odds in their mind that they can win, and that’s the fact of life. Analysts has to identify the chances or likelihood of win and identify the gap which can affect the result.

Still prediction is hard. Apart from Uncertainty, which is the default condition for predictions, there are two special conditions which the business of predictions challenging, Volatility & Time Horizon

Volatility & Time Horizon

Volatility is uncertainty in action. Things are happening so fast that past trends and patterns have evaporated. Normalcy and business as usual are things of the past. Now the contingency plans are out in the open and survival is name of the game. No one is sure of the future, be those managing from inside or analyzing from outside. Crisis is looming though not arrived yet. Volatility is thus a period between the normal and crisis. After crisis comes the disaster, a fear on everyone’s mind.

Predictions and forecasting go haywire in periods of volatility. Most of the lazy analysis is carried on the premise that future will imitate the past. Not anymore. The analyst is out of his/her wits and unwittingly resort to the doomsday or comforting mode, depending on the default. Volatility analysis requires combination of insights, perception and experience. A keen eye can identify patterns within the noise, though the job is not easy.

Time horizon is the second issue with mots analyses. All of us have written those personal essays with questions like ‘how you see yourself in next 5/10 years’, and have cobbled together responses but in the heart of the heart we know that it to be all BS. We honestly don’t know and worse still we cannot know or imagine ourselves over longer time horizons. In development sector, I see plans with two years and five years future looks. Two years one can muster something about, but five years look very very far. To stretch the prediction beyond a year or so, is an exercise in finding ways in a jungle engulfed by fog.

To identify the future beyond a year or so requires a combination of robustness, rigor and imagination. It also requires minute mapping while predicting the contextual and contributory factors and determining likelihood of each. Think of the issues like climate change, terrorism, inflation, AI, and think of predicting the future with confidence and the effort it will take to narrow down the scenarios. The way to go is to identify the accurate indicators.

Indicator based forecasting

Predictions are hard but we have to start from somewhere. I would revert to the contextual and contributory model and add indicators to it to take the framework into futrue.

Indicators are quantitative and qualitative representations of the factors which are shaping our problem. Let’s take Pakistan’s policy problems like dollar crisis or circular debt. The dollar shortage is created when massive amounts of dollars are smuggled to Afghanistan, and that happens when there is political instability with payments are to be made. Now smuggling will be the contributory factor and political instability in Afghanistan will emerge as the contextual factors. The question now would what are the indicators for these two? Smuggling is hard to identify but one can look at the massive withdrawals from the exchange companies as one indicator or changing geo-strategic factors affecting stability.

Let’s take another chronic issue, the circular debt. What are the contributary and contextual factors? Electricity theft is the contributory factor which is shaped the context of lack of accessibility due to insurgency. Now when the insurgency is on the rise, we can assume that utility bills collection will suffer leading to more leakages and more burden on the government to meet the gap.

One of my seniors, who has rare talent of being good at the analysis game gave a very simple and interesting example of indicators in assessing political risk. If in a particular country, the political unrest is always triggered by the prices of essential commodities and in a particular year the wheat crop has been hurt due to famine. A perceptive analyst will predict political unrest in next few months. Again, the prediction beyond few months is not easy.

We have made some effort at developing a predictive framework. Things too remember are- being Realist, identifying Range & limit, finding what is Probable vs Possible, thinking in Odds, finding Indicators and stretching the pathway from present to future.

We started with the five key questions of analysis ;

1. What is the Problem?

2. Where are we?

3. How did we get here?

4. Where would we go from here in future?

5. How do we change the future?

We have covered first two questions about problem statement and situations analysis.

Now is the time to tackle the third one.

HOW DID WE GET HERE?

You have noted right. HERE is our point A. By now we should have known what our point A looks like. Situation analysis will cover the dimensions of Point-A and the place around the point A looks. The latter we called Contextual Analysis.

Let’s get back to some examples of problems and issues.

• 44 % of school age children are out of school in our country. This is despite the extra resources directed to the provinces who have the responsibility of education portfolio exclusively. The situation is worse in the case of girl children and certain regions where the out of school children is highest. These are the areas that have also received maximum funding, but the school enrolment remains low.

• Company’s sales are dropping and 30 % of customer have been lost. This is despite the fact that company is spending highest on advertising. The new products have also failed to make a difference. Even sharing the margins with retailers is not working.

• Climate change is getting worse every year. The weather pattern is changing, rains are low, and productivity is falling. There is risk that Indus basin which is the breadbasket of the country will turn to drought.

AVOID THE SOLUTION YET, PLEASE

The situation analysis can be as elaborate and broad as one wants it to be, but after some time it will become problem repetition. We wanted to know about the problem, and we have gotten to know more than enough. Now the question is how to get out of the mess. One convenient and lazy approach would be what I call Import a Solution.

All of us have heard about the uncles who are walking physicians. They hear you sneezing and will immediately scream, aah this is new strain of flu, take the medicine I have taken, and you will be alright. Or you are having trouble finding a job and they know someone who landed a job through volunteering, so you need to volunteer as well. Policy sphere is also abundant in such quacks who have a ready made solution up their sleeves.

Of course, Import a solution hardly works but walking physicians are never out of fashion. In fact, jumping directly from problem to solution is popular but approach. I call this approach lure of solution. Ours is the time of quick decision makers who have answers before even the question surfaces. Leaders, CEOs and executives look bad if they don’t have a string of solutions handy. We are also encouraged to be futuristic, and of course we know that past is another country.

True and accurate analysis is linking and visualizing the reality in past, present and future. Problem and situation analysis define our present. Solutions are designed to avoid repeating the present in future. But where does present come from?

Causation analysis is determining the factors which shaped the present and removal of these factors will change the future. When talking about the causation, the famous adage needs to be made a header or footer so that it is never out of sight.

You have guessed it right.

CORRELATION IS NOT CAUSATION

Correlation-Causation debate has led to countless memes. My favorite is this one.

I would start with a basic concept from physics i.e function- a relationship between the dependent and independent variables. We have to identify the dependent variable i.e out of school children, falling sales and climate change and then make two sets of fictitious or dummy equations, ideal and problem. These are dummy because there is no quantitative backing. The point here is to identify the variables at least.

First the key equation. What will be the function for children enrolment, company sales, and climate change.

children enrolment = f(schools, teachers, parents’ commitment, community, distance…)

Sales = f(product, marketing, distribution, quality assurnace…..)

Climate change =f (emissions, population….)

The above are the normal functions while the problem equation would have independent variables going in the opposite directions.

Low children enrolment = f(less schools, discouraging teachers, hostile community, unwilling parents)

Low Sales = f(bad product, weak marketing, dysfunctional distribution, ineffective quality assurance)

Bad climate = f (high emissions)

CONTRIBUTORY/CONTEXTUAL FACTORS TABLE

In case you find the function and equation appears daunting to you (it does appear to me), you can make a table of your own to identify the factors. But one should be clear to make a distinction between the contributory and contextual factors.

The distinction between the contributory and contextual factors is crucial in mapping the pathway to present problem. Too many cliches and stereotypes are always circulating when it comes to policy and business problems. The purpose of analysis is to identify the contributory factors while considering the contextual factors which have shaped these contributory factors.

As the above picture shows, the problem exists in the contributory factors sphere which are shaped by the contextual factors. Poverty is contextual factor which affect whether parents send their children to school or not, but many poor parents do send their children to school. What is the defining difference between the poor parents who send their children to school and others who don’t?

Contributory factors directly, immediately and exclusively affect the outcome which results in the persistence of problem. Contextual factors are too broad, too general and almost too independent to be dealt with directly. To focus disproportionally on the contextual factors leads to analysis being open ended and almost resignation in front of problems. After all, the poverty, conservatism and narrow mindedness will remain forever. To blame them is akin to shirk one’s responsibility and move on with life.

BECAUSE STATEMNT

Causation analysis should conclude with the BECAUSE statement.

• Social media is anti-social BECAUSE……contributory factors first and contextual factors later

• 44 % of school going children are out of school BECAUSE …..contributory factors first and contextual factors later

• Company sales are dropping BECAUSE…..

• Climate change situation is getting worse BECAUSE…….

• Parliamentary system is better than presidential BECASUE……

• Economic crisis is crisis of public sector BECAUSE……

• The history will repeat itself in Afghanistan BECAUSE…..

• Leaders are made not born BECAUSE…...

• Crisis is only failed public relations BECAUSE……

  If you are not using BECAUSE, you are not exploring causation

Data Visualization- Trends and Patterns

Trends and Patterns are favorite tools of analysts. Even a pedestrian or basic analysis will appear good with some graphs thrown in and trend lines shown rising or falling.

• Trends emerge across time- Trend is relationship of one variable with time being the other variable. By showing trend, you only get to see what the problem looked back in time- was it less, more or constant. Trend thus show correlation, but we have to find causation.

• Patterns are formed with in a space. Patterns take time as constant and identify the space while mapping the problem. Patterns are territorial or geographical representation of the problem space.

Trends and patterns are useful & convenient but can also be tricky and deceptive specially when it comes to causation and forecasting. In-fact , trends and patterns hardly identify the relationship of one variable with other variable. Thus trends and patterns suffer from correlation-causation fallacy

Let’s look at some of these trends and patterns

This graph is classic Trend graph. It can be presented as analysis but to call it analysis would be no less than insult to the term analysis. It just shows the high level of reserves in Jan 22 and their rapid fall over thirteen-month period. Analytical question is finding the factors (contributory & contextual) which caused that fall.

The above shape is the classic pattern analysis electricity theft across districts in Pakistan. Red is above 90 % theft, brown between 80 & 90 & and white less than 5 %. It says as much it does - the rogue districts lie in Balochistan and KP Provinces. But why is the question that reader has to answer himself.

Taking the case of out of school children, here is one pattern of where the out of school children is found

Again, the map can be presented as the analysis, but it is merely mapping of problem. Analysis as we keep on harping, is finding how did we get here. For that we have to go deep and identify the specific patterns and then find the reasons behind the problem.

This one is better graph in so much that it includes both trend and patterns. The graph shows US assistance to Pakistan divided between economic, military and development aid. You get to see the periods of highs and lows of aid and also the composition or pattern. It is a neat situation analysis, but causation is still absent. Not that it cannot be deciphered. Like the informed observers of national security and international observers can identify the contributory factors like US intervention in Afghanistan and Cold War/War on Terror as contextual factors. But that is separate from this graph.

HISTORY TRAP

How did we get here, requires going into the past to find why we have the problem in present but final word of caution is in order. Past can also mean infinity just as future. How far in the past one needs to go. There is always a trap laid out which I call history trap- going way and way back.

Let’s take the chronic policy problems like poverty, unemployment, extremism, illiteracy etc. One can trace these as far as back to 1947 or even farther to the roots of colonialism. We are poor/illiterate/radical/unemployed because we were colonized, is the favorite trop in developing countries. I call it trope not because there is no historical basis to the adverse effects of colonialism. In fact there are more than sufficient evidence. My issue is with the reasoning which assumes a linear and straight causation ignoring all that is happening now, all the new developments and all the changes people have undergone recently and the world which is changing every passing day.

Apart from policy problems, the history trap is always lurking in personal, business and social problems. Just like the couples trace their issues to the very first day of marriage and the managers remember the day of interview when they were hired. In-fact, both would have seen high and low days since the first day. The history trap is something serious analysts should avoid, and solution lies in finding the inflection point.

Inflection point is the point when the tide started turning and did not recover. Like in the graphs shown above, one can see the inflection point when the foreign exchange reserves started falling and did not quite recover. The identification of inflection point depends upon analysts own insights but also on the data which is available to the analysts. A correct inflection point can give useful indications about the contributory and contextual factors.

Logically, the change in dependent variables is triggered by change in independent variables or factors. In case of sudden turn of events like fall in sales or foreign reserves, the factors that were sustaining the constant trend would have been replaced by the disruptive factors. In other cases, like out of school children, probably the fall would be more constant and slower. Still in that case, we have our inflection point given i.e the 18th amendment which was supposed provide positive contributory factors, but it does not appear to be the case.

Our inflection point will thus be to the left of our point-A. and the picture from the section WHAT IS THE PROBLEM, will look something like this

This part covers the area between the inflection point and problem point i.e point A. One of the common mistakes in analysis is start right with the inflection point or start in the past. That would be mistake because we can find or locate the inflection point with reference to the point-A. Otherwise, our inflection point will be arbitrary as is seen in many analyses which randomly pick a date in past and start telling the story.

The story part brings us to the writing part. The mapping of difference between the inflection points and point-A and identification of contributory and contextual factors will necessitate a story of what, how and why. Here the risk is that story will start looking like conclusion. In-fact the conclusion will come after the part-4 & 5 - Where should we go from here, and how to get there - solution. At this stage we only have what can be called as findings on the persistence of the problem. That is why the alternate name for this whole part can also be.

WHY THE PROBLEM PERSISTS?

I would recommend this question as sub-part because I consider process to be more important than the answers. In-fact better process will lead us to better answers. Otherwise, the temptation to jump to quick and shortcut answers is almost irresistible.

Taking the example of out of school children, if the question WHY THE PROBLEM PERSISTS is asked, there will be many short cut - conventional wisdom- readymade answers like poverty, radicalization, lack of infra-structures. These readymade reasons will lead to readymade solutions and the readymade solutions will have the same results as before.

But the answer is required to question WHY THE PROBLEM PERSIST and after mapping the journey, identification of factors and role of each variable into the persistence of problem, the analysis will reach a point where a definite and specific answer can be retrieved on the persistence of problem. This answer will be presented as the finding and will constitute a major milestone in analysis.

FINDING REASONS THROUGH REASONING

AI has a short and crisp definition of reasoning, “the process of using logic and evidence to think through situation and draw a conclusion. Exactly the sort of exercise we are indulging in. Identification and mapping of contextual vs contributory factors, drawing of trends and patterns and finding of outliers will provide us with the necessary evidence.

But how to apply logic.

One can fill pages and pages in explaining logic and it will remain elusive, that is the challenge of logic. So better to make it simple and close to the discussion about the problems and solutions, and our key question why the problem persists. Logic will help us getting to good cogent answers. How?

All of us start with some intelligence guesses and somethings accepted as truths. The intelligent guesses are called as assumptions and accepted truths are premises. We have to alos remember that we are in the realm of deductive reasoning i.e going from general to specific. Most of the analyses are based on deductive reasoning because the inductive exercise in reasoning is time-consuming, resource intensive and highly complex. Logic is needed along with evidence because there can never be enough evidence and answers come from choosing the best option through a rigorous and consistent process.

We can start with guesses which or assumptions and then employ evidence to test these assumptions. We are also aware of shortcuts called as premises. You might recall the famous example about premise- All humans are mortals, Socrates is human, therefore Socrates is mortal. The expression THEREFORE donates the operation of premise making. The fatal error in logic and consequently in analysis is to have incorrect, wrong or irrelevant premises and untested assumptions accepted as evidence.

Some examples of assumptions and premises in policy

• Employees are motivated by monetary incentives (premise) therefore increasing salary will lead to better performance

• People vote on the basis of quantitative indicators (premise) therefore if government is able to lower inflation rate, they will win the next elections.

• Children drop out of school because their parents want them (premise) therefore giving money to the parents will improve school enrolment

• Big governments are bad for the economy (premise) therefore slashing the bureaucracy will result in better governance

We can go on and on but the point perhaps clear by now is to answer the main question, how did we get here and sub-question why the problem persists. Answering these questions will lead us to the prediction about the future and devise measures to change the future.

More on that in coming pieces.

WRITING THE ANALYSIS

Doing the analysis and writing the analysis are simultaneous actions. The final product of analysis is in the written form therefore we should be mindful how we put the analysis on paper. Though the thinking part still has primacy, writing can elucidate or muddle the thinking. Writing can also be a barrier in a way that it comes too easy. Just take your pen or keyboard and start. But just writing is just the surface- of what is out there. And what is out there is also known to everyone so the analysis will remain superficial and at worst pedestrian. More on the fusion of style and substance later.

The best and most suited form for the analysis is essay. Modern knowledge specially in social sciences has been conceived and conveyed as essays. An essay is a fluent, coherent and self-contained form of composition. The opposites of essay would be a Report which as the name suggests is narration of ‘what happened’. The requirement of flow and coherence means that analysis is presented a composition of self-contained paragraphs without headings and numbering.

Writing aids like bullet points, graphs, tables and pictures can be used when best suited but aids should not dominate the narration. Sometime analysis contains table after table or graph after graph, needlessly and continuously leading to loss of meaning. One should not forget the purpose of analysis; to help make better decisions, strategy and policies. At the receiving end of analysis is a real individual with real problems who need to be directly talked too. I am personally a fan of more visual approach to understanding but only when narration is handicapped.

Analysis in written form can be called as Analytical essay or Analysis papers or simply the ANALYSIS. Analysis papers should not be confused with academic papers which are published in peer reviewed journals and are meant for academic audiences with more focus on how the analysis is done then the actual outcomes. Though the purposes of knowledge and analysis are different, even when analysis is presented as research. In both government and private sector, research departments or sections produce research reports and papers, though that is analysis in its basic form, even when their standard does equal or surpass academic research. Think tanks like RAND and private labs like that of IBM and Bell have produced research that has shaped academic discourse as well.

Academic research and analysis differ in purpose too, the former is to create knowledge while later is to solve problems. Research is essentially informed by past; analysis is meant to inform the future. Research is theory-making; analysis is policy making. Both have overlapping parts though when it comes to collection of data, literature review and analysis of that data to draw conclusions. But the key difference of clients and purpose should be kept on mind. Research is for fellow academics; analysis is for decision makers. At the same time general process and ethical conditions apply to analysis as well i.e data collection should be proper, results should reflect the data, and most importantly plagiarism is a cardinal sin in both.

Now back to our Question No. 2

WHERE ARE WE?

We remember the first question, what is the problem. We also defined the problem as difference between two points A & B. This question defines the point A and what is around that point as a place and time is concerned. Here is good place to introduce dimensions.

One dimension is straight line, and two-dimensional form is depicted in the form of horizontal and vertical lines. Three dimensional is most interesting. I keep on hearing and reading the word multi-dimensional in serious discussions and always wonder how can we capture three dimensions? Well, we can but it requires either advance mathematical equations or programs.

Also, how many dimensions points are out there?

A dimension by definition is a series of points with a direction and relationship between the points. A single point has zero dimension.

In analysis we start with a single point and find what is around that in terms of two dimensions, time and place. The idea of time and place will also help in data visualization. As already mentioned earlier, time and space come together to form situation. Situation analysis is the basic form or level of analysis, as seen in journalistic pieces and government reports which are commonly called as Situtaioners, also claiming to be analysis.

Actually defining situation is just start of the analysis and exploring what is around the point of problem in terms of time and place even when it is approached in terms of Ws like what, when, where and who. In-fact the human mind in my opinion only operates in two dimensions, and most minds in one direction only. Thats the tragedy whose consequence we see daily.

What will you do if given or having made your problem statement? What does anyone do when face with a situation requiring analysis? How will you start ?

We google and nowadays we go to ChatGPT.

But what is wrong with that.

Everything, I would argue.

Not that it is lazy, it can also be misleading. World would be simpler if all the answers were on Google. You can get answers to question what foods to eat when you have diarrhea or how much is the distance between London and Paris, what is the number of out of school of children in India. But you cannot solve problems and make decisions based on google. Google or internet will answer your queries, for solutions to complex problems, you are on your own.

The first question then would be-What does point ‘A’ looks like with reference to the problem statement we have made. where does point-A exist and what kind of environment shapes it. To leave the Problem point and jump to conclusions is the mistake we see every day but still cannot resist it.

The shortcut is to first identify time and space questions. What is the extent of problem at this time and how far back in time you have to look? How wide you go to find answers? and in case of really deep analysis, how much deeper you would go to retrieve information about the point on the surface.

Let us get back to the problem we are using as case study i.e out of school children.

Pakistan has 44 % of children out of school despite the spending on education steadily increasing post 18th amendment and the subject of education devolved to the provinces.

What do we want to know in terms of time and space. There is common saying about three set of queries: known knowns, known unknowns, and unknown unknowns. Here we are looking for the first known knowns, which is essentially someone else’s work already out there. As we progress, we will try to find the known unknowns-something new, and the third we might never know- the actual out of box stuff.

In known knowns arising out of time and space, I want to add more complexity with introduction a tool here -cause & effect. What we have in front of us is effect - 44% children being out of school and what we are looking for, is cause. If just add WHY anywhere in the statement, it will become cause and effect question.

How do organizations understand and manage cause and effect?

Organizations se including governments see the problem in terms of inputs and outputs, effect as output and input as cause. More schools are inputs and children in these schools are output. The problem we are having that a large number of children are not in schools. If our problem was the knowledge gap or learning outcomes, the input-output equation would have been different. In terms of looking at time and space questions with having inputs and outputs in mind, my situation analysis would include answer to following questions.

• What would be starting point of problem.

  • Here we are talking about 18th amendment, but we need to go a decade back to see if the trends are same or worsened. The first question thus would be;

    • What is the trend of out of school children since year 2000.

    Other ancillary questions would be;

• Out of school children proportion in provinces and further, as percentage of population, segregated by gender.

• Number of available schools and enrollment therein vis-a-vis the provinces and even further down

• Number of teachers in the available schools with corresponding enrolment

• Density of population and coverage of schools (to see if distance is the problem)

• Income level of the parents of out of school children

• Family size of the out of school children

• Education levels of parents of out of school children

You can go on and on. But the point here is to collect as much information as possible about the situation around point-A. You can take another major problem and find questions around that; low tax to GDP ratio, women in labor force, illegal migration etc etc. Another temptation here is to not immediately go for literature review. This is yet another difference between the analysis and academic. Analysis is about asking questions first, research is about finding gaps in knowledge, therefore you immediately go for existing knowledge in academic research. We would go the literature but later and with the point of comparing and testing the existing explanations.

MAKE YOUR OWN TABLE

Most of the information out there is presented in tables and tables are different version of either time and space both or further elaboration of either of the two. To make analysis fresh and different, convert your situation analysis into a table. Put the time in vertical axis and different parts of the space on horizontal axis. Lets take a look at the table for our given problem

Here we have time (years) on the vertical axis and different dimensions of the problems and the factors which could have contributed to the problem are on horizontal axis. You can add as many more as you find necessary and relevant. For more & bigger data, you can use Excel as well. But the point here is that rather copy paste the tables from others work, you create your own version of reality with the factors that align with your thesis.

SITUATION VS THE CONTEXT

Context is popular in analysis. when your analysis is critiqued for having missed key factor, you will retort by claiming that critique is out of context. On the other hand, if the reviewer inquiries about presence of seemingly irrelevant things in analysis, you will justify by mentioning the context. Context is critical in comparisons. When you would be comparing one thing with another and challenged by fellow analysts “but that is a different context” they might exclaim. You should have an answer ready for that.

In my understanding, situation and context distinct and differ by degree of intimacy with the problem or our Point-A. Situation is most immediate as indicated by our questions we have framed earlier. I like to use the word double-click since we are talking about one point. Keep on double clicking until it does not go further, and you will have clearer and clearer idea of the situation.

Context on the other hand is broader and donates the environment in which the situation exists. PESTEL (Political, economic, social, technological and legal) form of analysis is popular, and is example of explaining the context. It might not necessarily explain the problem or even the situation, but it will explain the environment in which the problem exists. Similarly the tools like SWOT provide context shaping framework.

Final word of caution. The necessary and relevant test applies and the law for that is “correlation is not causation”. Two things can happen simultaneously at the same place and same time but change in one might or might not cause the change in other. You might see that one province is ruled by political party and other by another one for more than one decade and be tempted to attribute the difference to political factors. That can be misleading shortcut unless proven directly.

How do we establish the context?

When the contextual factors become contributory factors (more on contextual and contributory factors in next parts). Climate change is one clear example. It has always been a contextual factor but now translating into contributory factor in all sorts of issues. Sindh province had Floods a decade ago-people were displaced in large numbers-mostly the displaced ones were of low incomes levels- displaced lived in camps for long time- camp had no schools- more out of school children. Here we see clear link of contextual factors contributing to the problem.

Again, you can take a look at big contextual factors like extremism, political instability, cultural conservatism etc and find the link with seemingly unrelated problems. But the word of caution remains- there has to be clear link, longer but still clearer. Her logic can play role, and logical fallacies can result in bringing down the whole edifice.

These contextual factors also have a critical role in finding solutions that they help us form assumptions. You start with assumption that low enrolment in schools is due to conflict which, but you get the data and find that enrolment is unaffected. Here the assumption does not hold. So many assumptions we hold dear actually do not hold when faced with test of analysis.

Anyways, context is most misused and least understood word in thinking and analysis as you can fit anything under its garb and leave anything. But without context the understanding and analysis are both incomplete. Taking the example of dimensions, I think the context is to have a three-dimensional view of the other two dimensions-also called as balcony view.

SITUATION ANALYSIS IN WRTING

In writing, the situation analysis should not take more then 20 percent of the piece. In case of one-pager, it has to be just one para and in five-page paper, one page. Since lot of data and information surfaces in situation analysis and one feels like throwing everything at the reader, selection suitable information is critical. There is another risk that situation analysis becomes endless. Most of the analysis I come across are just situation analysis, which go on and one in defining what is happening (situation) and what is happening around what is happening (context). The key is to follow the necessary and relevant test and include only the information and material which has bearing on our problem.

Now after defining the problem and situation, we are ready to move on to the defining question- How Did We Get Here?

Wait for the next part of series Analysis Made Easy.

Also check the other parts in the series Analysis Made Easy

What is a Problem after all?

To put it simply, problem is a gap. Yes, it is as simple as that. And if it is not clear enough, Problem is gap between two points (let’s call them A & B). Point A is where you are and point B is where you want to be.

This definition assumes the problem to be measurable. It also assumes that you know where you are and where you want to be.

Let’s start with simple ones.

Your body temperature is 103 C. It should be 98.6. The problem is clear, its the difference between the two , commonly called as fever.

Your income is 45,000 and your expenditures are minimum 90,000. The difference of 45,000 is the problem.

You are driving on the motorway and your car breakdown. The difference between your driving and the car not working is the problem.

You are married. Marriage means happiness to you but you are not happy. The difference between your current state(unhappiness) and where you want to be (happiness) is a problem. Though there is big caveat i.e you should know point B , which you might not be able to in most of cases. This does not mean that Gap model does not apply but this I call point B Problems. Most of the problems persist because you are solving the problem i.e bridging the gap but do not know where you want to be. Mostly human relations are point-B problems.

An employer is not happy with employees but the company as per employees is doing well. They attribute employer’s behavior to negativity and cynicism. Employer thinks that staff should do more and work hard but doesn’t tell them about his point B and everyone ends up dis-satisfied.

In other situations, the Point-B is set unrealistically.

A colleague of mine and his wife are both working. When he is back from office, he wants to lie down, watch tv and relax. But his point-B for his wife is that after work she should also do cooking, look after children and tend to his ailing mother. He does not agree and insists that problem is with her for not doing as expected (his Point-B). She argues back that it’s an unfair and unrealistic expectation (her point-B is difference and involves both of them).

A software company CEO wants his products to be rival Microsoft, while the company is new, small and in-experiences. He keeps berating the employees for being useless. He even offers them raises. But the product does not improve. The CEO thinks employees are useless. They regard him as having unrealistic expectations. Point-B problem persists.

PROBLE IDENTIFICATION’

Problem identification is when you are able to realistically identify points A & B and able to measure or map difference between the two. The measurement or mapping might not be quantitative. It might even indicate a feeling or state about which the parties involved are in agreement.

Many would argue that problem identification and definition constitute 80-90 percent of analysis. I tend to agree with the view that correct problem identification pretty much shape the analysis and save it from the D word.

Yes, to be called Descriptive, is a curse for the analysts. Description is echoing what the readers and consumers of analysis already know. Even reporting is better than Description. Most of the news we read in news are reports. We read them because they are fresh. Most of the news leave analysis to readers and at best provide a context. Even that is fine too. But to keep on describing the same thing over and over again will put you in bottom 10 percent if you are writing college essays and tossing of your paper in the bin if you are writing professional. It will also result loss of subscribers if you are on Substack.

Back to the problem definition. Why we start with the problem right away?

DEDUCTION, INDUCTION & HYBRID MODEL

You might have heard the difference between inductive and deductive reasoning or methodology. It is very hard to differentiate between the two. Commonly and conveniently, one is going from specific to general (inductive) or from general to specific (deductive). More elaborately in context of the research, if you start with a theory and there is already sufficient literature on the subject you go the inductive way. On the other hand, if you are starting without any theory- which is either something really new or the research is at large scale, you go for deduction.

As you have guessed, it becomes mind boggling and confusing to distinguish between deduction and induction. I use more convenient and self-explanatory terms for myself- convergence and divergence. Induction is divergence - you go from specific to general and deduction is convergence- from general to specific. Think in terms of one starting point and expanding it further- divergence or starting from one broad spectrum and bringing it to a single point- convergence.

Another question? How does problem fare in induction/deduction or divergence/convergence. I have got a convenient answer. Induction or divergence starts with problem and deduction, or convergence ends with problem. An example is in order here.

You are studying whether a particular drug is effective to control obesity. You start with a problem and a corresponding theory and diverge to find whether the drug works or not. On the other hand, you don’ what causes obesity, and which drugs works, you start broadly-with a general problem and converge to a specific problem- what causes obesity. In the former, you confirm the solution and for latter you confirm the exact nature of problem.

Another convenience is in order. Just forget the deduction as a standalone. When both terms are used together one assumes that both are common. Actually, they are not. Most of the research and analysis are inductions i.e you start with one theory, or a problem and you expand your work, or a combination of induction and deduction which i will call hybrid approach. As in the case of example of obesity, finding the causes of obesity will require a project covering large population and spread over number of years. Not something individual researchers or analysts expected to do. And we are not discussing research as such but analysis (the difference is of depth though the fundamentals of research still apply). So for most part think in terms of induction or divergence.

Another question needs to be addressed straight away. If we are going the induction way and use the divergence approach, how would we reach a conclusion?

This is the key question and is indeed what happens. I conduct five analytical writing workshops every year attended by 50 civil servants and out of 250 analysis papers I read, just a handful have conclusion borne out of analysis. In most cases, conclusion is missing and in the rest is abrupt and forced. The scheme of analysis become something like this.

Here we are starting with one point (problem) and then diverging and then keep on diverging. Conclusion will be lost in this expansion and then the analysis will end with conclusion is dropped from above and will appear detached and forced.

The solution lies in our hybrid model. You start from single point with problem, expand and diverge and then at some point in the middle start converge and end with another single point- conclusion. The whole process will take the shape of horizontal diamond.

Finally, we are settled on the Diamond model and now is the time to address the problem which while writing will come as problem question or problem statement.

Why the proportion of out of school of children in Pakistan is 44 %, second highest in the world ?

or

44 % of total children in Pakistan are out of school, second highest in the world.

The first is problem question and second is problem statement.

Again I would argue that one needs a blend of both to start a holistic and substantive analysis. Also the analysts should add value to problem statement on his/her own part. Just to say that 44 % children are out of school is what made the analyst turn towards the problem. The problem is still not identified and that will require some work.

I propose that problem statement has three parts.

Statement of fact — The problem question—Statement of intent

To say that 44 % of children are out of school is just a statement of fact. There is still no question. We can always add ‘why’ before the statement and turn it into an analytical question. After all, why question is the ultimate question. Why 44% of children are out of school, should suffice for analysis. I would hesitate to do that despite being in agreement in principle. First, an open ended why question is a lot of work and the scope will not be clear, and second thesis is absent from the statement.

Thesis or hypothesis is the version of reality we know already. In this case of out of school children, we know that there is some explanation already out there. There should be lot of work on the problem and a plethora of explanations on why the problem persists. The analysis does not operate on tabla rusa or clean slate, but has to cater the un-answered questions out there and incorporate the hypothesis out there. Not because hypothesis is fancy word, but to remain focused and build on the foundations of work out there.

The questions can be any one or combination of three.

•Why is this analysis required?

•Why is the problem persisting?

•What's the missing link(s)?

How about giving more context and create a question.

Pakistan has 44 % of children out of school despite the spending on education steadily increasing post 18th amendment and the subject of education devolved to the provinces.

Now there is question within the statement. The spending is increasing, and the responsibility is now with the provincial governments. Why is this problem of out of school children persisting then, is the problem question. Here we can insert our original framing of problem being difference between Points A & B. Point A is 44 % children out of school. What should be the point-B?

We are again confronted with Point-B problem which is common in public policy. Here we have exposed another big problem. Despite the jargon and clamor, we are not really following the ideas like Result Based Management (RBM) or the Rate of Returns type approaches. May be someone might go to old files and retrieve that we had committed 25 % children out of school with existing resources.

Now we are getting somewhere. We can bring the out of school children to zero but that will require more resources than we can muster. For now we can only work with what we have. The question, the problem question is that why we are failing in bringing the out of school children down to a reasonable level given all the investments we have made in the public education.

We have now covered two parts of the problem statement, the statement of fact and problem questions. The third part remains , the statement of intent. What do want to do in this analysis ? What will be the outcomes ? What are we trying to achieve here.

Many analysis papers a separate section of scope but this is good place to indicate the scope. The statement of intent will be added to our existing problem statement something like this ;

Pakistan has 44 % of children out of school despite the spending on education steadily increasing post 18th amendment and the subject of education devolved to the provinces. Policy, resource distribution within education and implementation factors needs to be studied to identify reasons behind high and rising ratio of out of school children.

Now we have added scope and intent to the statement and indicated the possible outcome, which is analysis of policy decisions, resource distribution and implementation issues. Why we have chosen the three things in scope, is also not with reason. Policy, resources and implementation are three tools available with the government. Government makes policy-location, type & size of schools, distribute resources- buildings, transport, food and then does implementation-attendance, monitoring and evaluation. Hopefully through study of combination of factors, we will come up with reasons behind high out of school children ratio and our solution will focus on removal of these reasons.

This is also good place to plug in definition of problem definition from my favorite work of literature

Problem is something you know. What you don’t know is why the problem persists, and how you are going to fix that.

(Zen & the Art of M-cycle maintenance p-131)

The problem statement is something you know (the fact), exploring why the problem persists (problem questions) and how it will be fixed (solutions)

PROBLEMS, ISSUES & ISSUE MAP

The difference between issue and problem needs to be addressed at the onset too as both are used interchangeably. I think there are two ways of looking at it.

One, issues are about people and problems are about systems. In the example quoted above in which husband and wife have difference on the expected role of wife who is also a working woman, is a question of issue not problem.

At more social level, valentine day is an issue in our society. It might not be a real problem. But at the same time, as along as it is issue between the people, it is a problem.

The other way of looking at issue/problem dichotomy is to place issues below the problem. You have problem at the top but within the problem, there are always layers and these layers are issues. Like in the case of husband and working woman wife, there are issues of culture, family system and also simple and plain misogyny.

Same is the case with the example we are using as case study i.e out of school children. There are number of issues which shape and define the problem at the onset. These issues should inform our analysis and lead us to more clarity on why the reality persists and how can it be altered.

When we locate the issues within the problem, the better tool is to have an Issue Map. The issue map is similar to the idea of mind-maps with just the difference that while you throw everything in the mind maps, issue maps are more focused. Lets look at the basic issue map of our problem of out of school children.

Here we see the issues with problem at the center. We have issues like less schools, disproportionate emphasis on buildings, religious and cultural factors, poverty and opportunity costs attached with sending children to school. Then there are always issues within the issues. We can also link one issues with another and sub-issue arising out of major issue as the girl’s education is an issue which is attached with sub-issues like role of community, culture and some districts will be more prominent than the other in girl children.

NECESSARY & RELEVANT ISSUES

Issue mapping is the tricky part in analysis as prone to the cognitive biases. If its a chronic problem, people have surely formed opinions about it and these opinions are now quite hardened to become biases.

Lets continue with our case study of out of school children. There will always be a group which will advocate for more infra-structure. Less schools is the issue. This group will be countered by the culture group. It is all about culture, they will argue. Then some will define the problem in terms of better monitoring of teachers and absence of KPIs.

Who will take precedence in this?

I think the resolution of which issues to consider lies in locating necessary and relevance issues. For this I will propose an intimacy test. How close the issues to the problem apparently. It has to be apparently because we are just starting the analysis. We will be sure once we are done with the work, but we cannot explore every issue and sub-issue for the sake of time, focus and energy.

The intimacy test would result in identification of necessary and relevant issues. It requires little logical work.

Continuing with the out of school children case study. The problem is with out of school children. Who makes the decision to send children to school? Its the parents. Now the issues which affect the decision of parents to not send children to school, is necessary in our analysis.

But parents are one side of picture. In policy its called demand side of problem. There is supply side too, schools and the paraphernalia. Thats the necessary issue as well. And problem persists due to combination of both demand and supply side issues; parents are not sending children to study and schools are either not there or are not receptive.

Then the issues which further affect the necessary issues come in the category of relevant issues. Why parents are not keen in sending children ? May be its poverty, culture or community. Or the schools are far, don’t have washrooms, teachers are not forthcoming, there is no accountability.

As one can see, the policy problems like out of school children will have layers upon layers of necessary and relevant issues. The key is identification and locating. If we start with the school building or with the teachers monitoring right away, we would be ignoring the necessary issues and jumping right to the solution. A fatal error in analysis.

We have just covered the first question of analysis in the series Analysis Made Easy. But it is a major milestone. More episodes of the series to follow.

Analysis is not something being taught at universities. Business, investment, political and security are academic and professional subjects. Those who study and practice might be called as experts.

But analysis?

How does one do/perform analysis?

I asked many analysts but didn’t get a cogent answer. Yet so many analysts are out there churning out analysis after analysis. Governments, businesses, media, intelligence, academia; analysis is everywhere.

So, I embarked upon the quest to explore analysis on my own and found that most of analysis can be pretty much summed up in five simple questions.

These are the five questions.

Question 1; what is the problem?

Everything starts with problem. If there is no problem, there is no need of analysis. We can resort to gossip and anecdotes, surprise each other with fun facts and even react to events around us. But these would be what they are, anecdotes, gossip and feelings. Not analysis.

Problem can be a statement- our sales are falling, or a question- why our sales are falling?

or another statement- Climate change is causing disasters, or a question- what to do with climate change causing disasters.

Problem can be just an issue (more on difference between problems and issues later), is AI good or bad for education?

or peek into future problem, how would Trump presidency look like for immigrants?

Question 2; Where are we?

A little too simple but I would start with the description of the present, with respect to the first question. If you have to give your problem or issue under discussion a shape, what would that shape look like, a circle, a block, or a line. I like to start with a point, just one point, and what is around that point as a place and point in time.

Time and space come together to form situation. Situation analysis is the basic form or level of analysis. Mostly journalists publish situation reports-description of events at certain time and place. They also have their own hacks for these reports like the Ws- what, where, when, who etc. Investigators, intelligence agencies government commission similar reports. These reports are termed as analysis, but I would call these as what they are, just reports.

In terms of analysis of issues or problems which are more out there, like social media, politics, war, fashion, economy, there would still be some basic information. This would form the answer to where are we. If I am thinking of my issue as one point or one dot, I would like to describe that very dot and then what is around. The word of around the point is familiar but fundamental. It would be called context.

Context is not just the description of the place but the relationship of that particular point with the time and place. You would like to remember context because when you would be comparing one thing with another and challenged by fellow analysts “but that is a different context” they might exclaim. You should have an answer ready for that.

The point, what is around (situation) and what is around the around (context) would hopefully give us a picture of reality.

But is that analysis?

Surely not. What next then?

Question3; How did we get here?

Simple question with a complex answer. This is the point from where the real fun starts. Till now we have identified where we are re the issue, and the situation and then the context around that. We have defined the reality. Now it is the time to go deeper and ask where this reality came from, or how did we get here. This would require good old-fashioned reasoning

Reasoning takes number of forms ; going into cause and effect, relationships between variables, causation vs correlation, immediate vs distant factors, relevant vs irrelevant and so on

Lets take some examples

Why did sales of our product drop ? ……Is our product bad? Is it the new competitor? Have the tastes changed or is it due to government regulations?

Why is climate change suddenly causing disasters? ……Is it due to reaching of the inflection point? Is it fruit of industrial revolution? Is it man-made or natural?

What won Trump the Presidency of United States?…..Is it due to the commitment of white voters? Is it due to disillusionment from the liberal crowd? Has the silent majority risen against the vocal minority?

Finding reasons makes the analysis worth its name. But still, what is the use of delving into past? What use are the old stories if we are still blindsided about future ? what good it will bring if we know how we ended poor, sick or miserable?

Yes indeed. That is why two questions remain.

Question 4; where will we go from here?

Analysis is not complete without forecasting. In-fact analysis is the study of past to predict the future. In my understanding forecasting starts with one fundamental assumption. Will future look like the present?

If you have studied Bayesian framework, you will be better informed than I am to have peek at the future based on priors. But I will answer the simple question still, will the future look like present and if not, how?

Here you will enter into two possibilities.

• Future will be like present. Company sales will keep on dropping, climate change will be worse, and Trump will be unstoppable. or

• Future will be different from past. Company sales will pick up, climate change will stabilize, and Trump will make mistakes and become a different person.

Which option to choose?

The task is simple without analysis. Optimists will choose the second one and pessimists will choose the first one, or other way around depending upon the issue.

What will analyst do?

Analysts will try to resolve some of the complexity by developing a system of indicators or parameters which will give him certain confidence to identify a future trend or pattern (just to be sure, trend is about time and patterns are about space). To arrive at the relevant and necessary indicators is the real skill of the analyst. He or she might get help from data and computing, but analysis is not throwing in the data right into the face of reader, analysis is about giving meaning to data.

Good indicators will produce good analysis. How to develop good indicators? Wait for the subsequent posts. For now, we are concerned about the next stage, and we have established that analysis should show a future informed by past and present.

But is that enough?

So far, our analyst is a passive observer- finding, exploring, predicting reality. What about changing the reality. Companies, governments, intelligence agencies and think tanks do not hire analysts to tell them fate accompli. They want the answer to existential question.

Question 5; What would it take to change the future?

Bold and uncertain territory. But we have come here after answering the Question No.4 and have made certain predictions about future. We have also identified certain parameters, certain levers that are driving the future reality.

Now it’s time to think about changing the future. In-fact any change minor or major is about changing the future. we think we are changing us in present but actually present is when the decision is made. Action is all in future.

Analysis concludes with these actions.

Analyst’s client is interested to know, what would it take to boost company sales, prevent the climate change disasters or give a stop to the plans by Trump and co.

The catch is that most of the analyses would already be talking about these actions, or solutions. But if you have answered or tried Questions 1-4; you would know where you are, what is around that point, how did you get here and what parameters would shape the future. Then you can identify the openings from where change can be instituted. That would be one superior analysis, but that is hard work.

This is all about the Analysis- 1. More in the part-2