The forecast that survives one bad week.

A forecast can be accurate and still be fragile.

Here is the number that should bother you. When a single deal carries more than 40 percent of a quarter, it fails to close about four times out of five. So a forecast that hangs on one big deal is not a strong forecast. It is a coin you are afraid to flip.

We spend all our energy on the wrong question. We ask, is the number accurate. The better question is quieter and meaner. Does the number survive one deal slipping.

The most accurate forecast I ever backed still cost us a quarter, and it was not even my number on the spreadsheet.

The branch sales manager owned the branch number. The national sales manager owned the region. I owned it for the whole organization, and I had given the company my word we would land this account. A large healthcare deal, bigger on its own than the branch’s entire quarter, five reps’ quotas in a single signature, and the boost the company was already counting on.

Everything we controlled was ready. The product. A new truck. A new service rep assigned. Install scheduled for the last month of the quarter. On paper the number was booked, and I had told the organization as much.

Then the call came. The customer had told their current supplier they were leaving and given thirty days notice. Their contract required sixty. They had misread their own agreement, so they could not take our service for sixty more days. Not a big deal, you would think. Except that extra thirty days pushed the install out of our quarter and into the next one.

We kept the customer. We missed the quarter. The branch and the region came up short on revenue and on sales target, the company held budget but without the boost everyone had been promised, and every cost we had frontloaded, the truck, the rep, the product, sat another thirty days with nothing coming in against it.

The forecast was not wrong. The account was real and it closed. The number just could not survive one thing moving, and the one thing that moved was a clause in the customer’s contract that nobody, including me, had thought to verify. I had staked the organization’s expectations on a deal I never stress-tested.

Why does an accurate forecast still blow up?

Because accuracy only measures whether you called the total. It says nothing about what the total is standing on.

Take two forecasts. Both commit the same number. One is built on eight deals that each carry a slice. The other is built on one whale and seven maybes. On the sheet, they are identical. In a bad week, they are not remotely the same animal. The first one bends when a deal moves. The second one breaks.

Warren Buffett said it best about risk. “Only when the tide goes out do you discover who’s been swimming naked.” A bad week is the tide going out. It does not care how accurate your spreadsheet looked in July.

So stop grading the forecast on accuracy alone. Start grading it on survivability. Four tests, in the order they matter:

1. Isolate what the number rests on. Pull the deals that make up your committed number and find the biggest one as a percent of the whole. This is the single-deal exposure test, and it takes four minutes. If any one deal is more than 40 percent of the quarter, you are carrying a passenger you cannot afford to lose. That is not a forecast. That is a hope with a due date.
2. Grade coverage on quality, not just size. Three times pipeline sounds safe until you look at where it came from. Close rates are not equal by entry point. Expansion and add-on business closes about three out of four. A referral closes about one in five. Cold outbound closes closer to one in a hundred. So three times pipeline built on cold outbound is not three times pipeline. It is a rounding error wearing a suit. Weight your coverage by how the deals actually entered, not by the raw dollar count.
3. Run the Tuesday question on every load-bearing deal. For each deal holding up the number, ask one thing: what has to be true by Friday for this to still be real. If you can answer it with a specific event, a signed order form, a scheduled security review, a budget confirmed in writing, it is a commit. If the answer is a feeling, it is a hope. Sort the whole forecast that way and watch how much of the number moves from one column to the other. And inspect what you expect. On the healthcare account, the thing that had to be true was that the customer was actually free to leave their old supplier on time. We assumed they had read their own contract. They had not. If it can push your install date, it belongs on the list, even when it lives on the other side of the table. Verifying it is not nagging. It is protecting the number for both of you.
4. Report survivability up, not just a confidence percent. When your leader asks about the quarter, do not hand them a number and a percentage. Hand them the number, the one deal it depends on most, and what happens to the quarter if that deal slips a week. That is a leader talking. A confidence percent is a rep guessing with more decimal places.

The spike and the system.

An accurate forecast is a spike. You were right once. Good for you. A survivable forecast is a system. It holds up when a deal moves, and deals always move. The rep who gets lucky on a whale looks like a genius for one quarter. The rep who builds a survivable number looks steady for ten years. Only one of them is still carrying a bag at the end.

The short version.

1. Accuracy asks if you called the number. Survivability asks if the number can take a hit.
2. Run the single-deal exposure test: any one deal over 40 percent of the quarter fails to close about four times out of five.
3. Weight coverage by entry point. Expansion closes about three in four, referral about one in five, cold outbound about one in a hundred.
4. Use the Tuesday question. If a deal’s next step is a feeling, it is a hope, not a commit.
5. Report what the number rests on, not a confidence percent.

Here is your homework: pull up your committed number and find the one deal it cannot survive losing. Then answer the only question that matters about that deal, which is what has to be true by Friday for it to still be real. If the answer is a feeling, you do not have a forecast yet. You have a hope with a due date.

This essay ran as edition 11 of The Frontline Sales Forecast, the weekly newsletter. One issue a week, published here on the Friday, then emailed and posted to LinkedIn the same morning. Same piece everywhere. Get the next one.