"The market is worth six billion dollars. We only need one percent."
We have deleted that sentence from more business plans than any other. It is not a forecast. It is a hope with a decimal point.
What bottom-up actually means
A bottom-up forecast builds revenue from the mechanics of your own business: how customers find you, how many convert, what they pay, and how long they stay. Every number in the chain is something you can measure, argue about and improve.
Step one: the acquisition engine
Start with how customers arrive. List each channel separately, because they behave completely differently:
- Paid acquisition, with spend, cost per lead, and a conversion rate
- Outbound, with number of reps, activity per rep, and a meeting-to-close rate
- Organic and referral, usually a function of your existing customer base
- Partnerships, modelled per partner, not as a blended assumption
Each channel needs its own volume assumption and its own conversion rate. Blending them hides the fact that one channel is subsidising three others.
Step two: pricing, segmented
If you have more than one customer type, forecast them separately. A model that uses a single blended average revenue per customer cannot show what happens when the mix shifts, and the mix always shifts.
Step three: retention, from cohorts
This is where a forecast becomes credible. Take your actual customers, grouped by the month they joined, and track what proportion remain each month afterwards. That curve is your retention assumption. It is real data about your real business.
If you do not have enough history yet, say so explicitly and use a conservative benchmark with the source noted. An investor will forgive a young company for lacking data. They will not forgive an unmarked guess.
Step four: assemble
Revenue in any month is then: customers carried forward, plus new customers acquired that month, minus churn, multiplied by segment pricing. Stack it by cohort and the picture builds itself.
The sanity checks
Before you show it to anyone, test three things.
Capacity. Does your forecast require more salespeople, support staff or production than the cost model funds? Growth has a cost line.
Implied share. Calculate what percentage of the market your year-five revenue represents. If it is implausible, the forecast is wrong regardless of how carefully it was built.
Reasonableness of change. If your model shows conversion doubling with no change in product or process, delete it.
Why this matters
A bottom-up forecast usually produces a smaller number than a top-down one. Founders resist it for exactly that reason.
But the smaller number comes with an answer to every question an investor can ask about it. That is what gets funded: not the size of the projection, but the strength of the reasoning underneath it.