Research
Markin Research: first-party work on B2C growth decisions
A recurring line of research into the decisions that move ARPU and margin. Each report declares its sample, its period, the definitions that change the answer, its sources and its limits, so it can be re-run rather than trusted.
We publish a number only when it came from a randomised holdout or from arithmetic a reader can rebuild. Where a finding rests on anonymised Markin deployments, it is labelled as such and carries no aggregate figure, because the sample is too small and too varied for an average to be honest.
Every report is reviewed twice a year, and re-run out of cycle whenever a source it rests on changes. Written by Romà Llambés, CEO, Markin. Every report is also published as plain markdown at the same address plus /md.
How this research is produced
- What is Markin Research?
- A recurring editorial line of first-party analysis on B2C growth decisions: when a retention discount pays, which decisions carry margin, what a company needs to run per-customer decisions, and what survives when ARPU and churn are defined consistently.
- Where do the numbers come from?
- Two places, always labelled. Reproducible arithmetic over public information, where the method section lets you rebuild the result yourself. And anonymised ranges across Markin deployments, used only where public information cannot answer the question, with sample and period declared.
- Why do some findings have no number attached?
- Because the sample behind them is small and the businesses are very different, so a mean would be a rounding of something rather than a finding. Where that is the case we publish the mechanism and refuse the figure, which is the same standard we apply to other vendors' claims.
- Can I cite these reports?
- Yes. Every report has a stable URL, a plain-text twin at the same path with /md appended, a declared review date and a full source list. If a figure is past its review date it is removed rather than assumed to still hold.