I Gave Myself 15 Minutes to Tell a CFO If We Were Going to Hit Plan

The Google Data Analytics Professional Certificate has done a few people dirty with how it structured the capstone.

You get through seven modules, you get to the capstone, and it hands you two prepackaged case studies and a shrug. Do the bike share one. Do the wellness one. Or go figure out your own project.

Almost everybody does the bike share one. Which makes the bike share study the Titanic dataset of that program. Everybody has it in their portfolio, and exactly zero people care about it.

So this is the other option. The one they leave you alone on.

The rules I gave myself

Fifteen minutes. Obnoxiously short, on purpose, because the whole point is reps.

If you can go soup to nuts on a case study in fifteen minutes, think about how many datasets you could get through. How many different business situations you could sit in. You get dangerous a lot faster doing twenty of these than doing one big fancy one that falls on deaf ears.

All of it in Excel. No Python, no SQL. Not even AI, because these are my stripes to earn. If you want to use those tools, go for it. Excel is my bread and butter because everybody already has it, most of the business world runs on it, and honestly it is a lot more capable than people give it credit for these days.

One disclosure. I wrote the requirements for this dataset, but I did not build it and I had not opened it. So the mess in it was a surprise to me too.

The situation

You are an analyst at a private equity rollup.

There have been a few acquisitions, and a couple of those companies came in with their own CRM. Salesforce, HubSpot, whatever. Somebody smashed them together into one instance, the way this always happens, with about the success rate you would expect.

The first budget just came down from the board. You had no say in it.

Then the CFO sends you a file and derails your next fifteen minutes:

Can you tell me if we are going to hit our bookings target this year?

Three tabs. CRM data, invoice history, and the bookings plan. We are halfway through the year, and because things have been going well enough and everybody has been heads down on acquisitions, nobody has actually looked forward yet.

Clock running.

What was in the file

The CRM tab was opportunities. Product names with a weird prefix on the front, ENT1, ENT2, ENT3, which is the entity each deal came from. Booking amounts. Create dates going back to 2024 against close dates all in 2026, so a long lead time. Stages: closed won, closed lost, discovery, negotiation, proposal, qualified, verbal.

The invoice tab was the problem.

Those product names do not look like the products on the other tab. Not close. There are hundreds of them. It is messy, and I hate it, and it is so normal that hating it is beside the point.

The bookings plan was clean and small. Four categories, by quarter. A full year target of $63.8 million.

Start calculating before you know what you are looking for

This is the part I would tell anybody starting out.

Staring at a fresh dataset is writer's block. You are looking at a blank sheet of paper thinking you cannot come up with anything. So do not try to plan the analysis. Just start doing calculations and be curious. The shape of the answer shows up on its own.

I started counting.

700 closed won. 1,150 closed lost. 1,850 total, for a 37.8% win rate. Nice round numbers, suspiciously so, which is what a built dataset looks like.

Then the same thing in dollars with SUMIFS, and the dollar win rate came out within half a point of the count based one. That is worth noticing and it took ten seconds. It means average deal sizes are consistent between the deals you win and the deals you lose, so you are not winning a pile of small ones and losing the whales.

What is still open

Closed deals only ran through July, which means August through December is still live.

So I added a helper column: an IF that returns N when the stage is closed won or closed lost, and Y for everything else. Open or not open, one column, and now COUNTIFS and SUMIFS can slice on it.

970 opportunities still open.

Banked so far: $37.8 million against a full year plan of $63.8 million. Against the first half target of $28.7 million, that is roughly 9.7% ahead year to date, which is where a lot of people would stop and report good news.

The gap shows up

Here is where it turns.

Open pipeline times the historical win rate is about $18 million of expected additional bookings. Add that to the $37.8 million already banked and you land at roughly $56 million.

The plan is $63.8 million.

That is an $8 million gap. And to close an $8 million gap at a 38% win rate, you need somewhere around $20 million of new pipeline that does not exist yet.

Being 9.7% ahead year to date and being $8 million short on the year are both true at the same time. That is why the question the CFO asked is not the same as the question the year to date number answers.

How long a deal takes changes the answer

The gap on its own is not the answer. The gap plus the clock is.

Close date minus create date, averaged over the won deals: 80 days mean, 52 days median. Call it two months.

So the pipeline that closes this year has roughly two more months to get created. That reframes the whole thing. The question stops being whether we can find $20 million and becomes whether we can find it soon enough for it to land.

Breaking a personal rule

I do not like pivot tables. I have made videos about not liking pivot tables.

I built one anyway, because I needed pipeline creation by month and I had two minutes left.

Average of about $10.4 million of new pipeline per month. The last couple of months are soft, but they are the summer. October and November get big. This business books in the fall.

Two months of runway at roughly $10 million a month covers a $20 million need, if the seasonality holds and if the trend is real. I tried to split it by entity to separate real growth from mix, ran out of clock, and got general growth across the board rather than a clean answer. Which is its own finding.

Pencils down

Here is what I would say walking into that room.

If we stopped generating leads today and closed our existing pipeline at the 38% rate we have been closing at, we would add about $18 million. That takes us to $56 million against a $63.8 million plan, so we are looking at an $8 million gap. Covering it needs roughly $20 million of new pipeline at our current clip, and we generate about $10 million a month with our two strongest months ahead of us.

So it is achievable. It is also tight, and it is at risk. The thing I would do about it right now is make sure marketing has the resources to drive the leads, because that is the variable we can actually move.

That is an answer with the right level of caution, and it has one lever attached to it.

Now the caveats

What you just read is directional by design. It is data backed as well as it can be in fifteen minutes, and it comes with a pile of footnotes. Both of those things are true at the same time.

It is not a finished piece of work. If somebody asked me to refresh it in three weeks I would open the file and wonder what in the world I was doing, assuming I could find the file at all.

That is honestly what a lot of this work looks like. Nobody talks about it because it does not feel like great work.

The part that gets you noticed

But it is important work, and these small asks are seeds for the bigger work.

While the CFO is in that meeting, you get to slow down and ask your own questions. What does this look like at the product level? Are average deal sizes going up or down? Why don't we have a product map that spans the CRM into the invoice history?

And then the one that actually changes your job: why don't I just build one?

Those are the questions worth having. And acting on one of them, before anybody asks you to, is the thing that gets you noticed.