Build a workplan and get the data you need

How to turn a hypothesis into execution, what to do when you have no data, and how to pivot when the hypothesis breaks.

A good workplan keeps you aligned and focused. A bad one creates rework and confusion.

Start from evidence, not intuition

Build an issue tree for each part of the problem. What needs to be true for the hypothesis to hold, and what would make it fail? Then, for every branch, define the fact base you need. What facts, analyses and data points would actually answer the question?

Be brutally clear here. The more precise you are, the less time you waste later. This looks like a duplicate of your hypothesis and it is far more specific, because you are also naming data cuts, time horizons and formats. If you are unsure, try reciting the hypotheses out loud. "I believe, X to be true because ...".

Then define the data

Usually some combination of internal client data, external market data, benchmarks, expert interviews, and sometimes competitor financial disclosures. Think through the minimum set required to make a confident call.

Do not ask for everything. Large unfocused data requests slow down the team and frustrate the client, and they signal that you have not decided what matters.

Draft the request early and write it precisely: specific fields (revenue by branch; revenue and profitability by product), time periods (monthly, last 18 months), business units, and formats (units, currencies). Make it easy for the client team to understand and gather. Show a draft to your manager or a senior consultant before it goes out; they will sharpen it and spot the pitfalls. Think about the analysis and deduct what data you will need.

A bad data request: "Hi, please send over the sales data for the stores you own, I'd like to look into SKUs"

A good data request: "Hi, could you share the monthly sales by store (focus on the Boston Area) and SKU (separated in volume and avg. price) for the last two months in USD."

Scope the analyses and sequence them

What do you need to build? A cost baseline, a market model, a profitability bridge, a synergy case? Break each into concrete tasks and sub-tasks, estimate how long each takes, and identify the dependencies.

Some analyses need data you will not have until day three or day five, so sequence accordingly. Start with what you can already do and prepare your templates, so that when the data arrives you can move immediately.

Then package it into a simple workplan answering three questions per assertion:

•  What tasks and sub-tasks are required to substantiate this assertion?

•  How long will each take, when can it be completed given external factors, and by when does it need to be ready?

•  What is the key output for each task, and in what format will it be delivered?

And get every meeting into the calendar now. When do you need your supervisor? When do you present to the team? When do you need which expert? Align this with the client updates and internal check-ins so you are never caught off-guard.

The workplan checklist

1.  Does the workplan directly lead to answering the client’s key questions?

2.  Is there a task corresponding to every aspect of your hypothesis and its assertions?

3.  Have you estimated how long the tasks will take?

4.  Is the timeline feasible, with some buffer built in?

5.  Are you clear on how to get the resources and information required?

6.  Have you identified the critical path, the part that cannot be shifted or sped up?

7.  Are you clear on the output per task?

8.  Are all required meetings with your supervisor and the client in the calendar?

9.  Is the workplan still simple and high-level enough to actually manage and update?

That last one matters more than it sounds. As an associate, I regularly tried to over-deliver by building a thorough Excel workplan. After a full morning of work, I filed it somewhere and never opened it again. Wasted time. Put the key topics in an email or on a slide you can share with your supervisor and the client, so the format gets used and iterated collectively. Updates should take a few minutes a day, not a morning.

When you have no data at all

Some of the hardest moments in consulting are a blank page. New case, a workstream nobody has touched, no reports, no benchmarks, no expert contacts. This is where some associates freeze. Four moves:

Start with first principles. What must be true for this business to work? Sizing a market breaks into population × penetration × frequency × price. A cost structure breaks into inputs × volume × unit cost. Get the structure right first, then fill in numbers as you work them out.

Use analogies. If you know nothing about the target market, find a comparable one you understand. How does the bicycle market in France compare to Germany? On one case, we were looking at meat alternatives in Europe, where little data existed at the time. So we studied milk alternatives instead, soy and almond drinks, and extrapolated sizes and adoption curves as one data point.

Build the issue tree top-down. Question at the top. Break it into two to four sub-questions that would fully answer the parent. Break each down one more level. Stop where you can see what data or analysis would answer each branch.

Triangulate from three weak sources. You rarely get one perfect data point. An expert estimate, a back-of-the-envelope calculation and a dated industry report can together create a credible range. This is how most consulting numbers are born. The skill is combining imperfect sources intelligently rather than hunting for a perfect one.

When the hypothesis breaks

Your hypothesis will be wrong sometimes. That is how the method works. It is a tool to focus effort rather than a prediction you have to defend. The danger is staying wrong because you got attached to your original answer.

Change it when data consistently contradicts your assumptions, when experts independently point elsewhere, when your model needs increasingly heroic assumptions to work.

When you see those signals, do not quietly adjust a cell and hope. Surface it: “Our initial hypothesis was X. The data is pointing toward Y. Here is what I think we should test next.” That is what good problem solving looks like from the outside.

Then go back to the issue tree, identify which branch broke, reformulate that branch, and adjust the workplan. Keep everything that still holds. You almost never start from zero, and most pivots affect one or two branches rather than the whole structure.

START THIS WEEK

1.  Write down all analysis that is needed to crack your case and sequence them into a high level Workplan.

2. What data do you need for the full case? Write a concise and clear data reuqest.

3.  Put every meeting your workstream will need into the calendar today, including the ones three weeks out.

A great workplan is one of the most powerful ways to reduce stress and keep alignment. It is not project management theatre.