118% net retention. 78% logo retention. Which one is your business?

A net revenue retention of 118% is the line every founder wants on the cover page. It says the company would grow even if it never closed another deal. Existing customers expand faster than they leave, the base compounds on its own, and the sales team is pure upside. Investors pay up for it because it is the rarest thing in software: growth that doesn't depend on next quarter's pipeline.

The number is real. But what sits underneath it might not support the story it tells. NRR is a dollar-weighted ratio, and dollars can be concentrated. A strong headline can coexist with a business that is losing most of its customers, leaning on a few large accounts to cover the gap, and counting reactivated churned customers as if they never left. None of that shows up in the ratio. It shows up in the composition.

The gap between dollar retention and logo retention

NRR counts dollars. A single enterprise account that doubles its contract offsets dozens of small logos walking out the door, because the ratio weights by revenue, not by customer count. The headline says 118% net retention. The logo retention sitting next to it says 78%.

That gap is the first thing an underwriter looks for. The dollar figure says the existing base is healthy. The logo figure says the product is failing most of the customers who try it, and the franchise depends on a handful of accounts continuing to expand. Both numbers are correct. They describe different businesses, and only one of them is the business you are running.

When Levian builds growth accounting from your transaction data, it computes aggregate net dollar retention from the revenue breakdown. Separately, it rebuilds logo retention by acquisition cohort. Read together, those analyses expose the gap between the dollars the base retains and the customers it keeps, even though they appear in different sections of the report.

Concentration carrying the number

A high NRR driven by broad-based expansion across the customer base is a different animal from one carried by the top ten accounts. Revenue concentration data answers which one you have.

If the top ten customers account for 60% of revenue in a given month and those same accounts are expanding, the NRR headline reflects their behavior, not the base's. The long tail of customers could be flat or contracting and the ratio still clears 100%. The number is accurate. The signal is narrow.

Levian computes revenue concentration by rank bucket each month, top 1, top 2 to 5, top 6 to 10, and so on down to the tail, so you can see whether retention is a property of the whole base or a feature of the accounts at the top of the stack.

Resurrection quietly reflating the base

A customer churns in March and returns in September. Growth accounting classifies this correctly as resurrection: not new revenue, not expansion, but a previously lost customer re-entering the base. The five-bucket decomposition, new, expansion, contraction, churn, and resurrection, keeps these separate.

But most dashboards do not run growth accounting. The returning customer shows up as new revenue or expansion, and the original churn stays on the books. The net effect is that both sides of the retention equation look better than they are: churn was real, the return was real, but booking the return without connecting it to the departure flatters the ratio.

When resurrection is large relative to new revenue, it means the company is working hard to refill a base it already lost. The headline NRR may still clear 100%, but the underlying motion is a treadmill, not a flywheel. Separating resurrection from new revenue is what makes that visible.

What this means before the raise

An investor will not take a retention number at face value. They will put dollar retention next to logo retention and look for the gap. They will check whether the top accounts are carrying the expansion. And they will look for resurrection revenue hiding inside the new-business line. A founder who has already done that work is negotiating from the same page. A founder who has not is about to learn their composition in the room.

Levian builds aggregate net dollar retention from your transaction data, then gives you the other evidence needed to interrogate it: logo retention by cohort, concentration by rank bucket over time, and resurrection separated from new revenue. The views are distinct, but together they show whether the number on your cover page survives diligence.