The ROI Guarantee
If we don't make you money, you don't pay.
Fulcrum guarantees a positive return on investment. Once each 12-month contract term ends, we calculate the value of your holdout-tested lift - using the conservative, lower-bound end of that measurement - and compare it to what you paid. If it falls short, we refund the difference.
Measured by a randomized holdout test, using the lower bound of a 95% confidence interval. Illustrative 12-month figures.
The promise
You only pay Fulcrum when Fulcrum pays for itself.
Most conversion platforms hand you a tool and hope you find the ROI. Fulcrum takes that risk off your books. We run a permanent, randomized holdout on every account - a slice of your traffic that never sees personalization - and measure the difference in revenue per user between the two groups. If that measured lift, valued conservatively, doesn't exceed our fees over your contract term, we refund the shortfall. No dashboard estimate, no click-tracked running total. A controlled measurement, and a guarantee behind it.
How it works
Four steps. One promise.
-
We optimize
Our personalization campaigns go live across your funnel, tuned to how your visitors actually behave.
-
We measure
A small, randomized slice of your traffic - 10%, fixed for every client - never sees personalization. Everyone else does. We track conversion rate, average order value, and revenue per user for both groups continuously throughout your contract.
-
We add it up
At the end of your 12-month term, we calculate the lift in revenue per user between the two groups, using the conservative lower bound of that measurement rather than the raw number - then compare it to the fees you paid over the term.
-
We stand behind it
If your guaranteed value falls short of your fees, we refund the gap - and the refund never exceeds the fees you paid for the term.
How we measure
A controlled measurement, not a running tally.
We measure one thing for your account: the difference in revenue per user between shoppers who get Fulcrum's personalization and a small, randomized group who don't. That comparison - not a count of clicks and orders - is what your guarantee is based on. At the end of your 12-month term, we calculate that lift at a 95% confidence level and use the conservative, lower-bound end of the range as your guaranteed figure. A qualifying result means the lower bound of that range is greater than zero - in other words, we can be confident the lift is real, not statistical noise.
Holdout lift, lower-bound estimate, 95% confidence = your guaranteed value.
The holdout is the entire basis for this number - not a separate check on it.
Measured over your whole contract
One number, calculated once, at the end of your term.
We don't check in monthly and we don't stop early. Your holdout runs continuously for the full 12 months, and we calculate the lift exactly once - after the term ends, using every bit of data collected along the way. A longer measurement window means a more precise result, so there's no early snapshot that could be misleadingly good or bad. The number we calculate at the end is the one your guarantee is based on.
If we fall short
We refund the difference, and only the difference.
This isn't a full-refund gimmick. If your guaranteed value - the conservative, lower-bound lift figure - exceeds your fees, you pay for that value. If it comes up short, we refund the gap. The refund can never exceed the total fees you paid for the term.
Example: illustrative, 12-month contract
- Fees paid (12-month term)
- $60,000 (illustrative)
- Qualifying traffic over term
- 400,000 users (illustrative)
- Lower-bound RPU lift (95% CI)
- $0.1125 per user (illustrative)
- Guaranteed value (traffic × lower-bound lift)
- $45,000 (illustrative)
- We refund
- $15,000 (illustrative)
Questions
The details, straight.
Is this a full money-back guarantee?
No, and that is deliberate. We refund the shortfall: the difference between what you paid and your guaranteed value for the term. You always pay for value you actually received, and you are never out of pocket relative to that value.
What if you do great for six months, then have one weak month?
There's no month-to-month tracking to worry about. We measure once, at the end of the 12-month term, using everything collected across the whole period. A single slow week or a single great week doesn't move the number on its own - only the full-term data does.
How do I know the revenue is really from Fulcrum?
The guarantee is based on a randomized holdout test - a portion of your traffic that never sees personalization, running for your full contract term. We compare revenue per user between that group and everyone else, and use the conservative lower bound of that measured difference as your guaranteed figure. It's a controlled comparison, not a count of clicks and orders.
What exactly are you counting?
The difference in revenue per user between your holdout group and the group that gets personalization, measured across your full 12-month term at a 95% confidence level. We use the conservative, lower-bound end of that range - not the raw average.
Is it based on revenue or profit?
Revenue. The guaranteed figure is top-line revenue per user measured against a holdout group. Your margins, cost of goods and fulfillment are internal to your business and outside our control, so the guarantee is measured on what we can directly influence and cleanly measure.
Is there a cap on the refund?
The refund can never exceed the total fees you paid - it's the gap between fees and your guaranteed value, so the most you'd ever get back is what you paid.
What if my traffic is on the lower end?
You still get a guaranteed number, calculated exactly the same way as every other client. A smaller traffic base produces a wider confidence interval, so your guaranteed figure is simply more conservative - you're never told there isn't enough data for a number.
Put our fee on the line.
See what a measured, dollar-for-dollar ROI guarantee could do for your funnel.
Start Your Free Pilot