TOOLS/ROI & incrementality
What is your decisioning programme actually worth?
Most ROI calculators multiply your revenue by a lift the vendor chose. This one shows both numbers: what a before/after read would claim, and what is left once you subtract the share of decisioning programmes that independent research finds deliver no real lift. Then it tells you the holdout you would need to prove it.
The revenue the decision layer is allowed to move.
Accounts that generated revenue in the last 30 days. Not registered users.
Recurring plus non-recurring revenue divided by active customers.
Margin on the next unit sold, not blended company margin.
Verified annual impact
$2.5M
Net incremental gross margin in the central case, after the programme cost and after the share of decisioning programmes that independent research finds deliver no real lift.
Reported uplift
$7.8M
What a before/after dashboard would claim, with no control group.
Verified uplift
$5.4M
What survives a holdout in the central case.
Return on programme cost
3.7×
Payback
4 mo
If 20–40% of it does nothing
What it takes to prove it
To detect a 3% lift on revenue per customer you need roughly 40K customers in the control arm, about 4.4% of your addressable base, read over at least 8 weeks, so novelty is not mistaken for effect.
Addressable base
900K
Revenue at risk from churn
$145.7M
Annualised, at the current monthly rate.
How the number is built.
01
Addressable revenue, not total revenue
A decision layer can only move revenue it can reach. The model starts from the share of your base that is consented, non-fatigued and reachable in a given month, and applies the assumed lift only to that.
02
Reported uplift is shown, then discounted
The uncut figure appears next to the verified one on purpose. It is the number most business cases are built on, and the gap between the two is the point of the exercise.
03
Margin, not revenue, is the outcome
Incremental revenue is multiplied by gross margin on the next unit sold, then the fully loaded programme cost is subtracted. A programme that grows revenue and loses margin has not worked.
04
Every estimate carries its experiment
The control-arm size and the measurement window are computed from the same inputs. If the lift you assumed is not detectable on your base, the calculator says so instead of returning a number you could never prove.
The formulas
addressable revenue = customers × contactable % × ARPU × 12
reported uplift = addressable revenue × assumed lift
verified uplift = reported uplift × (1 − haircut), haircut 20–40%
net margin = verified uplift × gross margin − programme cost
control arm per group ≈ 16 × CV² ÷ lift², CV ≈ 1.5 for revenue per customer
Sizing uses a two-arm approximation at roughly 80% power and 95% confidence. It is a planning estimate, not a substitute for the power analysis you run against your own variance.
Evidence standard
Why we discount before we add.
BCG reports that when organisations adopt rigorous incrementality testing, they typically find 20% to 40% of their active next-best-action programmes deliver marginal to negative lift.
Independent researchBCG, How Measurement Is Evolving in Next-Best Action (2026)The same research flags novelty effects, new programmes show inflated early results, and recommends 8 to 12 weeks before drawing conclusions.
Independent researchBCG, How Measurement Is Evolving in Next-Best Action (2026)Global-holdout, programme-level ROI measurements often overstate impact through halo effects, pull-forward effects and experiment contamination.
Independent researchBCG, How Measurement Is Evolving in Next-Best Action (2026)- Every decision Markin makes carries a control group. Uplift is reported against that holdout, not against the customers who did not qualify.
- Results are read over a full measurement window rather than in the first weeks, so novelty is not mistaken for effect.
- Programmes that fail to beat control are retired automatically. Killing decisions that do not pay is part of the loop, not an annual review.
- The one figure we quote about ourselves is a range, not an average: +17% to +35% ARPU on treated cohorts against a randomised holdout, across Markin deployments in large B2C bases. We publish no industry benchmark, because we could not source one we would be willing to defend. Your holdout is the number that matters.
Questions people ask before they trust a number like this.
What is the difference between reported uplift and incremental uplift?
Reported uplift is the revenue observed after a programme launches, compared with before. Incremental uplift is the revenue that would not have occurred without it, measured against a randomised control group. Reported uplift absorbs seasonality, halo effects, pull-forward and customers who would have converted anyway, so it is almost always the larger of the two.
Why does the calculator subtract 20% to 40% before showing a result?
Because independent research says it should. BCG reports that when organisations adopt rigorous incrementality testing, they typically find 20% to 40% of their active next-best-action programmes deliver marginal to negative lift. Applying that band as a haircut produces a planning number you can defend in a business case, rather than the best case presented as the expected case.
How large does the control group need to be?
It depends on the effect you are trying to detect, not on a fixed percentage. Revenue per customer is heavily right-skewed, so detecting a small relative lift takes far more exposures than a conversion-rate test. The calculator sizes the control arm from your assumed lift and tells you when the effect is too small to be detectable on your base at all.
How long should a decisioning experiment run before you read it?
At least eight to twelve weeks. New programmes show inflated early results because of novelty effects, so a two-week read will overstate impact. The calculator never returns a measurement window shorter than eight weeks.
Does Markin publish an average uplift figure?
No. There is no independently verified aggregate uplift benchmark for this category, from any vendor, that we would be willing to defend. Your own holdout is the only number that matters, which is why this calculator outputs the experiment you would need to run alongside the financial estimate.