SaaS Financial Modelling and Cohort Revenue Specialist

Kishen Patel

Founder, SaaSFractionalCFO.co.uk | ICAEW Chartered Accountant

Kishen works with UK SaaS and AI companies as a fractional CFO, building cohort-based financial models that give founders, boards, and investors a clear view of retention, payback, and revenue quality. He has helped over 35 SaaS businesses raise capital and build investor-grade finance functions from the ground up.

LinkedIn Profile | ICAEW Verified | Secure Your Readiness Call

By Kishen Patel  |  April 2026  |  12 min read

How to Build a Cohort-Based SaaS Financial Model

If every customer looks the same in your model, your forecast will mislead you. A cohort-based SaaS financial model groups customers by the month they start paying, then tracks how each group performs over time. That matters because blended averages can hide weak retention, slow payback, or newer cohorts that are quietly improving.

For founders and finance leads, this changes decision-making. You get cleaner forecasts, stronger investor reporting, and a better grip on what growth is actually costing. If you are preparing for a fundraise or building your first investor-grade SaaS financial model, cohort structure is where to start.

---

Start with the Right Cohort Structure Before You Build Anything

A cohort model only works if the starting point is consistent. Once you choose the rule, keep it. If you change it later, the whole history stops being comparable, and you lose the ability to spot trends across cohorts.

For most subscription SaaS businesses, first payment is the best cohort start point. It marks the moment a customer becomes real revenue, which makes retention, MRR, payback, and cash flow easier to line up in a single model.

Choose the Event that Starts Each Cohort

You usually have three choices. Each can work, but each answers a slightly different question about your business. Use this table to pick the right one for your model.

If you run a free trial-heavy model, signup cohorts can help you study conversion and activation. If you bill annually before service starts, invoice cohorts may suit revenue planning better. For most growing SaaS firms, first-payment cohorts are the most practical choice because they connect directly to paying retention. The key rule: keep one definition across the entire model.

Use Monthly Time Buckets to Track Performance

Monthly buckets keep the model readable. You track month 0, month 1, month 2, and so on for each cohort. That simple layout makes retention curves, expansion patterns, and churn timing straightforward to spot.

Monthly cohorts also match how most SaaS firms report. MRR, payroll, ad spend, board packs, and management accounts usually run on a monthly cycle. Your model should match the rhythm of the business, not fight it.

---

Build the Core Tables that Make the Model Useful

A good cohort model is a chain of connected tables. One feeds the next. If the logic is clean, the forecast stays consistent. If the logic is messy, the numbers drift apart and trust drops quickly. Start with customers, then move to revenue, then margin. That order makes the model easier to audit and update each month.

Create a Customer Retention Table for Every Cohort

Begin with a table that shows, for each cohort month, how many customers started, how many churned, and how many are still active. The key point here: do not use one flat churn rate across the whole customer life.

Most SaaS businesses see higher churn in the early months, then a slower decline once customers settle in. A flat rate hides that pattern and often makes forecasts look smoother than they really are. A better approach is a retention curve. For example, a cohort might retain 90 per cent of customers after month 1, 84 per cent after month 2, 80 per cent after month 3, then decline more slowly after that. Those percentages can come from historical data and improve as your product and onboarding improve.

This table also helps you ask better questions. Did churn spike in the first month after a sales push? Did a plan change improve retention for later cohorts? You cannot answer that with one company-wide churn number. This is where cohort analysis earns its place in the model.

Add a Revenue Table for MRR, Expansion, Downgrade, and Churn

Once you know how many customers stay, move from customer counts to money. For each cohort, track starting MRR, retained MRR, expansion MRR, contraction MRR, and churned MRR. That shows whether revenue growth comes only from new sales, or whether existing customers are becoming more valuable over time.

This matters particularly now because many SaaS businesses combine subscription pricing with usage or seat-based expansion. A customer may stay active but still shrink in value. Another may add users and grow far beyond its starting contract. If your model ignores those movements, your forecast will miss both upside and downside risk.

You should also separate logo retention from revenue retention. A cohort may lose a few customers but still grow in value if the remaining accounts expand enough. That is why many high-growth B2B SaaS firms watch net revenue retention (NRR) closely. A result above 110 per cent is often seen as a strong sign at Series A and beyond, though it depends on stage and customer mix.

Include Gross Margin and Contribution by Cohort

Revenue alone can mislead. A large cohort with poor delivery economics can look healthy on paper while contributing less cash than a smaller, better-priced group.

Add gross margin by cohort after revenue. If your direct costs change by customer type, plan, or service level, reflect that in the model. This is especially useful for hybrid SaaS firms that include support, onboarding, or service-heavy implementation work.

When you track gross profit by cohort, you can see whether newer customers are genuinely better customers. Perhaps your ACV rose after a pricing change, but support costs rose too. Perhaps enterprise contracts look attractive, yet lower-touch self-serve customers produce better contribution margins. That view helps you avoid false confidence: growth with weak margins often creates cash pressure faster than the headline revenue suggests.

---

Layer in CAC and Payback so the Model Can Guide Spend

A cohort model should do more than describe history. It should help you decide where to spend next. That means linking customer outcomes to the cost of acquiring those customers in the first place. This is where many models fall short: they track cohort revenue carefully, then leave acquisition cost blended at company level.

Match Acquisition Spend to the Cohorts It Created

Tie sales and marketing spend to the month and segment that produced each cohort. If possible, split by channel, such as paid search, outbound, partner, or product-led sign-up. Without that match, payback analysis becomes guesswork.

A model that blends all CAC together can hide a serious problem. One channel may bring in customers with strong retention and healthy expansion. Another may look cheap at first, then churn badly within three months. The same rule applies to customer type: if your enterprise cohort comes from an account executive team and your self-serve cohort comes from paid social, those costs should not sit in the same bucket.

• Split CAC by channel (paid search, outbound, partner, product-led) wherever possible
• Keep enterprise and self-serve acquisition costs in separate buckets
• Tie spend to the calendar month it generated, not the month it was approved
• Recalculate payback whenever a channel’s retention profile changes materially

Calculate Payback Using Gross Margin, Not Revenue

Payback should reflect gross profit, not total contract value. If you spend £18,000 to win a cohort and recover £3,000 of gross profit per month, the payback period is six months. If you used revenue instead, you might think the cohort paid back faster than it really did.

Gross margin is what funds payroll, product development, and future growth. Revenue that carries heavy direct costs does not repay CAC at the same speed.

---

Make the Model Decision-Ready with Segments and Scenarios

Once the base model works, add only the layers that improve decisions. More tabs do not make a better model. Clear links, clean assumptions, and monthly updates do. Board packs, fundraising decks, and hiring plans all improve when the cohort model explains both growth and efficiency.

Split Cohorts by Channel or Customer Type When Patterns Differ

Segmentation helps when behaviour truly differs. Self-serve customers often churn differently from sales-led customers. SMEs often expand differently from enterprise accounts. Those differences deserve separate cohort tables. Still, do not split cohorts out of curiosity. Do it when retention, pricing, gross margin, or payback clearly changes by segment. Every extra cut adds maintenance work.

A good test: if a segment would lead you to a different budget, pricing decision, or hiring plan, model it separately.

Add Scenarios to Test Your Plan

Scenario planning turns a static model into a planning tool. Build a base case, then add upside and downside cases for new customer growth, churn, expansion, pricing, and headcount timing.

1. Base case: your most likely outcome given current trajectory and no major changes
2. Upside case: a plausible win, such as better activation rates or stronger pricing power
3. Downside case: common risks, such as slower conversion, weaker retention, or delayed hiring impact on sales capacity

This is particularly useful for cash runway and fundraising conversations. A small rise in early churn can pull down ARR fast. A modest lift in expansion can do the opposite. By flexing those assumptions, you can show your board or investors what happens under different outcomes without rebuilding the model from scratch each time. For support with your fundraising model, see our fundraising and investor readiness service.

---

Avoid the Mistakes that Make Cohort Models Hard to Trust

Most broken cohort models fail in familiar ways. They mix cohort definitions, use flat churn, ignore downgrades, forget gross margin, or fail to tie the cohort view back to total company revenue. That last point matters most.

• Never mix cohort definitions mid-model; pick one start event and keep it
• Replace flat churn with a retention curve based on actual historical data
• Include contraction and expansion MRR, not just new and churned MRR
• Always reconcile cohort revenue totals back to reported MRR or recognised revenue
• Keep the model simple enough that one person can update it monthly without breaking it

A model that nobody can maintain has no value, even if it looks impressive on day one. The goal is a working tool that improves with every monthly update, not a showcase that sits unused in a data room folder.

---

Summary: What a Strong Cohort Model Gives You

A cohort-based SaaS financial model gives you a clearer view of retention, revenue quality, payback, and cash needs. It replaces blended averages with something far more useful: a picture of how customer groups behave over time, and how that behaviour drives or drains your cash position.

Start simple. Pick one cohort rule, use monthly buckets, and build from customers to revenue to margin. Then improve the model as better data comes in. That discipline helps SaaS firms raise with more confidence, make stronger decisions when the stakes are real, and give their boards the financial visibility they are asking for.

If you would like help building or auditing your cohort model, our fractional CFO service covers this end to end. You can also read more about the SaaS metrics we track and report on for every client, or get in touch directly via the contact page.

---