Evidence
Real customer bases, rebuilt from transaction data
The revenue line is rarely what the headline suggests
Every engagement starts the same way. An anonymised transaction extract, no personally identifiable information, and a question the reported numbers cannot answer: which customers is this revenue actually resting on, and will they still be there in eighteen months?
These are anonymised reads from live work. Different sectors, different revenue models, different continents. The same method.
The case studies
Ecommerce · B2C · Europe
€1.92bn reported revenue. 5.4 million customers. The base could sustain €1,196.8m on its own.
Read the ecommerce customer base analysisFintech · Remittance · B2C · Pan-African
The raise was priced on $9.54m of fees. The customer base sustained $5.04m of it.
Read the fintech customer base analysisInsurance · Short-Term · Multi-Brand · UK
£682m of Gross Written Premium across five brands. 93% of customers who left one brand left the Group.
Read the insurance customer base analysisWhat is customer base analysis?
Customer base analysis rebuilds revenue from individual customer behaviour.
Rather than reading the revenue line and benchmarking a churn rate against it, every customer is scored on observed behaviour across every transaction, then tracked year on year to show who strengthened, who weakened, who left, and what revenue moved with them. That produces three things a reported number cannot: which customers the revenue depends on, how concentrated that dependency is, and whether the base is improving or already breaking before the numbers reflect it.
Keystone IQ calls the method the Revenue Quality Architecture™.
Three steps. Days not weeks. No PII required.
Access
An anonymised transaction extract from the portfolio company. No PII. No disruption to the business. We confirm that the data is usable before the engagement formally starts, so there are no surprises on scope or timeline.
Analyse
We work at individual customer level across every transaction, not cohort averages or modelled proxies. That produces a complete picture of which customers are actually driving the revenue, how concentrated that dependency is, and what has already changed that the reported numbers have not caught yet.
Extract
A clear commercial verdict: what revenue is structurally supported, what is already on weaker footing, and the single most important action before or after close. The output is legible to the operating partner, the deal team partner, the portfolio company CMO, and the portfolio company CFO.
The same methodology, now applied to diligence
- Used at entry to test whether the revenue line is strong enough to underwrite.
- During hold to track whether the thesis is playing out at customer level.
- At exit to build the customer durability story before a buyer's DD team asks for it.
See our services
FAQs
It is the assessment of the asset rather than the market. Commercial due diligence tells you what the market will do. Customer base analysis tells you what the customers already did, at individual customer level, and what that implies about whether the revenue line holds. It is evidence, not opinion.
Cohort analysis groups customers by when they arrived and reports the average. Averages hide movement in both directions at once. Keystone IQ works at individual customer level across every transaction, so a base where the best customers are leaving unnoticed and the weakest are growing does not read as stable.
No. The analysis runs on an anonymised transaction extract: customer ID, date, value and category. No names, no contact details, no personal data leaves the business.
Because data describes the past. Diligence is a question about the future. Someone who has run customer acquisition, retention and pricing inside a live business knows which signals are leading indicators and which are lagging artefacts. That judgement cannot be modelled. It has to be earned.
See what your customer base is really made of
Days not weeks. Fixed scope. No PII required.
Get in Touch