For management and day-to-day operations
What does 70 percent in a sales budget really mean?
How a sales budget can be checked against a company's own closed deals of past years, and what role an AI assistant plays in it.
Most sales budgets are not forecasts. They are negotiations between optimism and Excel. The data is there, and still budgeting can start with someone writing 40, 60 or 80 percent next to a sales opportunity. Such a number says little about what it is based on.
What was the starting point?
There were years of closed deals, spread across CRM exports and several Excel and CSV files. The question that occupied me: what does a number like 70 percent actually know?
A sales manager remembers the last few deals. The files contain years of results that no one can remember in full.
What was built?
I built the first version of this budget logic in Base44. It proved the idea but did not fit the way I work. Later, I worked with Lovable, Cursor and Codex.
It was not about yet another dashboard. It was about letting the current budget argue with its own history. So the budget should not only hold one number per opportunity but also show how well that number is backed by earlier deals.
How does the comparison with the history work?
A calculation layer looks for the closest usable historical comparison, for example by source, type of service, region or customer profile. It also shows the sample size behind the comparison.
This makes the number verifiable because it immediately shows how much data supports it. Is the 70 percent backed by fifty closed deals or by two? With two, I would treat the number as a hypothesis and not as a reliable close rate.
The same layer also finds:
- average sales cycles,
- months that are repeatedly weak,
- segments with a strong pipeline but weak conversion,
- customer profiles linked to larger or smaller projects.
What does this change about the budget?
A pipeline can look healthy over the year and still leave a gap in September. A deal at 70 percent can belong to a segment that has historically had a much lower close rate. A deal without a date can inflate the annual total without contributing anything to the months in which staff and capacity have to be planned.
For an engineering services company, such monthly gaps are not a side issue. They affect capacity, hiring, expertise and the question of which opportunities can be pursued responsibly.
How does the calculation deal with incomplete data?
Facts and assumptions stay separate. Where data is missing, dates, duration and monthly revenue can be estimated from hours, rates or the closest historical comparison. Each estimate keeps a note of the confidence with which it was made.
What role does the AI assistant have?
The assistant has a different job from the calculation. It helps me question assumptions, compare the model with what the company knows and explain why a number changed.
It does not know the future. It makes sure that intuition has to show what it is based on. It does not replace the sales manager. It adds one more voice to the budget discussion: the company’s own business memory.
What is the core in one sentence?
A sales budget becomes more reliable when every probability has to show which earlier deals it rests on and how many there were.
What question remains?
What would happen to your sales budget if every probability had to hold up against years of your own closed deals?