On This Page
The short version
- A price for an IDP company says how much its customers must pay in the years ahead. You can read that claim backwards from the multiple.
- At a 70% gross margin and a 10% cost of capital, 7x revenue means existing customers pay the same amount every year, forever, before sales and overhead costs. Anything above 7x is a bet on growth; our examples count only growth from existing customers.
- The US 10-year Treasury yield averaged 1.45% in 2021 and stood at 5.05% on 28 October 2026. For a company whose customers grow 5% a year, adding those 3.6 points to the buyer's rate turns an 8x price from three years of growth into 23.
- With three-year contracts, 23 years contain seven renewals. At each one the customer can compare your per-page price with what a language model charges.
- An EBIT multiple mixes what customers already pay with what new customers might pay. A buyer who applies it pays less for a company that spends on winning customers.
- Multi-year volume commitments put more of the needed growth under signed contracts. Retention data a buyer can check makes the remaining renewals easier to believe.
The question buyers answer
"What is the right price for an IDP company?" cannot be answered in the abstract, because every price is a forecast about customers. When Coupa agreed to buy Rossum in May, the price said something about how long Rossum's customers would keep paying and how much more they would pay each year.
A better question is what a given price assumes, and whether your customers can deliver it. You can work that out before you sit down with a buyer.
Read your multiple backwards
Take a company with a 70% gross margin, a round figure we use in every example here; use your own. Assume a buyer who wants a 10% return each year on what it pays. That required return is its cost of capital. For customers who pay the same every year and never leave, such a buyer pays 7x annual revenue: 70 cents of gross profit per dollar of revenue, divided by 10%. That is the value of a customer base that neither grows nor shrinks, before sales and overhead costs.
Every other multiple can be read the same way. Each one implies a net revenue retention that the buyer must assume forever: the revenue this year from last year's customers, divided by what they paid last year.
At 3x revenue, the price assumes customers shrink by about a tenth each year. At 10x, they grow by almost 3% a year, forever. At a 13.6% cost of capital, 7x means existing customers grow by 3.3% a year, forever. The same offer asks more of your customers when the buyer's rate is higher.
Higher rates, lower expectations
The US 10-year Treasury yield averaged 1.45% across 2021 and stood at 5.05% on 28 October 2026. A buyer's cost of capital is set above that yield. In the examples below we assume a buyer who used 10% in 2021 and adds the full 3.6 points, which gives 13.6% today. It is an example; buyers set their rates deal by deal. A higher rate makes later money worth less: at 2% a dollar due in ten years is worth 82 cents today, at 10% it is worth 39 cents.
The rate is one reason. Our German series on priced-in years (in German) held the cost of capital constant and still found that investors priced fewer years of growth into listed software stocks in 2026 than before.
UiPath, a listed US automation vendor with its own IDP product, is one example of slowing growth from existing customers. Its dollar-based net retention rate, as reported in its fiscal 2022 and fiscal 2026 year-end releases, went from 145% to 107%.
UiPath's existing customers still spend more each year, but the increase has been smaller every year since fiscal 2022. A buyer who extends that line expects the increase to keep shrinking.
The rate trap
Take two companies. Company A sells at 10x revenue and its existing customers spend 15% more each year. Company B sells at 8x and its customers spend 5% more each year. We assume that growth continues for some years and then stops, after which customers pay the same every year. The question is how many years of growth the price needs.
Company A needs about twice as many years at 13.6% as at 10%, from under three to about six. Company B needs almost eight times as many, from 3 to 23. Above a cost of capital of about 14.2%, Company B's customers would be worth less than 8x revenue even if they kept growing 5% a year forever. No number of years is then enough. Because 8x is close to that limit, small changes in the rate or in the 5% move Company B's 23 years a lot.
A company whose customers grow more slowly than the buyer's cost of capital has a ceiling: the value of its customer base if today's growth rate held forever. When the rate rises, the ceiling falls, and a price that was within reach in 2021 can end up above it.
Here is the same 8x offer split into what existing customers already pay and the growth the buyer must believe in.
At 10%, one eighth of the price rests on growth. At 13.6%, more than a third does.
The renewal count
IDP is sold on contracts that renew. We use three-year terms in the example. With one-year contracts the count is three times higher. A CEO can check a number of renewals against the company's own renewal history in a way that a number of years cannot be checked.
At 10%, Company B's price is earned within the first contract term. At 13.6%, renewals fall in years 3, 6, 9 and so on up to year 21: seven in a row, and at each one customers must agree to pay more than before.
Each renewal is a chance for the customer to compare prices. AWS lists Textract form extraction at $50 per 1,000 pages. In a February 2025 benchmark, Google's Gemini 2.0 Flash converted about 6,000 PDF pages to text per dollar, roughly 17 cents per 1,000 pages. The two are different services and neither is an IDP vendor's contract price. The gap between them is about 300 times. If your per-page price halves at a renewal, the customer must send twice the pages for your revenue to stay flat.
Growth is two numbers
A buyer pays for two kinds of growth: existing customers who pay more each year, and new customers you have not yet won. A higher rate lowers both, by different amounts.
Take a customer who pays 100 a year and whose spend shrinks 10% a year, at a 70% gross margin. At a 10% cost of capital that customer is worth 315 to a buyer. At 13.6% it is worth 267, down 15%. Now take a customer with the same spending pattern who has not been won yet, and assume winning it costs 140, about 17 months of its revenue. Its value is 175 at 10% and 127 at 13.6%, down 27%. The cost of winning is paid at the start and does not shrink with the rate, so what is left falls faster. The larger that cost relative to what the customer is worth, the larger the gap.
| Value at 10% | Value at 13.6% | Change | |
|---|---|---|---|
| Existing customer | 315 | 267 | −15% |
| New customer, 140 to win | 175 | 127 | −27% |
Per 100 of annual revenue, 70% gross margin, spend shrinking 10% a year. Our calculation.
Value from customers you already have falls less when rates rise than value from customers you still need to win. Reporting growth as one number leaves the buyer to guess the split.
The EBIT trap
Some buyers quote a multiple of EBIT, profit before interest and tax. That multiple also says how fast profit must grow forever. At a 10% cost of capital and 25% tax, and if all profit after tax is paid out, 7.5x EBIT assumes no growth at all and 10x assumes 2.5% a year.
EBIT counts the cost of winning new customers as an expense today. Take two companies with the same revenue and the same existing customers. One spends 15 of every 100 in revenue on winning new customers, the other spends nothing, so the second reports 15 more in EBIT. If a buyer applies the same EBIT multiple to both, the second company gets the higher price, although only the first is adding customers. Revenue from existing and new customers, reported separately, shows the difference that EBIT hides.
The margin question
Customers on long contracts pay per-page prices agreed before the model prices shown above. If a vendor has moved its own processing to cheaper models, its gross margin on those contracts should rise until they come up for renewal.
UiPath's numbers do not show that. Its GAAP gross margin stayed between 81% and 85% from fiscal 2022 to fiscal 2026, according to its year-end releases for fiscal 2023, fiscal 2024 and fiscal 2025 and the two linked above, and was no higher in fiscal 2026 than in fiscal 2024. UiPath does not report IDP separately, so its margin mixes document processing with everything else it sells. The published numbers give no sign of such a rise. Your own data answers the question for your company: how much of your processing runs on cheaper models today, and when do the contracts priced before that switch renew?
What to do in the next 12 months
| Lever | What changes | Effect on the count |
|---|---|---|
| Trade price for term | A lower per-page price in exchange for a longer term and a volume commitment | More of the needed growth is signed before the next renewal |
| Show retention evidence | Renewal history for every customer, with dates and amounts | A buyer can check whether the renewals the price needs are likely |
| Report cohorts | Revenue by the year each customer signed | Shows how spending develops after each renewal |
| Split growth in two | Existing customers and new customers, reported separately | Shows a buyer which kind of growth it is paying for |
| Work out your own count | Your multiple, margin, retention and contract length | You know what an offer assumes before the buyer explains it |
A lower price for more volume only pays if the extra pages cost little to process. If processing costs rise with pages, a 20% price cut for 30% more pages raises revenue by 4% and lowers gross profit at a 70% margin.
Retention moves the value of the existing base quickly. At a 10% cost of capital and a 70% gross margin, one point of net retention, from 100% to 101%, raises that value by 12%. From 95% to 96%, it raises it by 8%.
How to work out your own count
You need four numbers: the multiple on the table, your gross margin, your net revenue retention, and the cost of capital the buyer is likely to use. Divide gross margin by the cost of capital. If the multiple is below the result, the price assumes your customers shrink. If it is above, the difference is the part of the price that rests on growth, the orange part of the 8x chart above. Your finance team can turn it into years with the formula below, then divide the years by your contract length.
For your finance team: with gross margin m, cost of capital i, net retention n, multiple V and q = n / (1 + i), set C = m·q / (1 − q). The years of growth are ln[(V − C) / (m/i − C)] / ln q. When n is below 1 + i, C is the ceiling, and if V is above C no number of years is enough.
Caveats
All valuations here are our own calculations under stated assumptions. They count gross profit as the value a customer brings and leave out the ongoing cost of serving and selling to existing customers, which lowers every figure. The cost of winning a new customer in the growth section is an assumption, not a measured value. The model assumes customer growth continues at a constant rate for a number of years and then stays flat. Real buyers use more detailed models, and their cost of capital differs by deal and buyer. The 10% and 13.6% rates are examples; only the 3.6-point rise in the Treasury yield is measured. UiPath figures come from its SEC filings. Page prices are a cloud provider's list price and a third-party benchmark, not prices customers pay.
Sources cited in this post: FRED (Federal Reserve Bank of St. Louis), UiPath 8-K filings with the SEC, AWS Textract pricing, Gigazine, christopher-helm.com.
We do not sell IDP software. We are not paid by any vendor mentioned in this post.