# Reading retention *Pull the number apart before you trust it.* --- You open the board pack. Net revenue retention: 115%. The benchmark is 100%, and you're comfortably above it. The existing base is growing on its own, before the sales team signs a single new deal. [[The churn ceiling]] explains why this number matters. 115% looks like a business compounding nicely. Pull the gross retention. --- Gross retention strips out expansion and measures what you kept of the revenue you started with. The board pack doesn't show it. 82%. Nearly a fifth of your starting revenue either left or spent less. The 115% is real, but it's 82% kept plus 33% grown from the survivors. The spread tells you how much the headline depends on expansion. --- Now look at who's leaving. Non-renewal isolates the decision. Of the customers who came up for renewal this period, how many chose not to? It's [[Switching costs|revealed preference]]: what customers do when they have the option to walk away. Your non-renewal rate is 18%. Nearly one in five customers who could have left, did. But 18% only counts customers who reached a decision point. Half your enterprise base is on three-year contracts and hasn't faced the question yet. Report 95% annual retention on three-year terms and actual renewal is closer to 85%. The retention rate measures the base, including customers who couldn't leave if they wanted to. The renewal rate measures the decision. Different numbers, different stories. --- Now look at what's inside that 33% growth. You pushed a 10% price increase in Q2. Customers absorbed it. That's roughly a third of your expansion, and it has nothing to do with product adoption. Customers paid more for the same thing. Seat growth and usage make up another chunk. Some is organic adoption, but some is the customer's own business growing underneath you. They didn't choose more of your product. Their headcount grew and your per-seat pricing grew with it. The rest is cross-sell and upsell: new modules, tier upgrades. Customers actively choosing to buy more. But one enterprise account tripled its contract after a divisional rollout, and removing it drops net retention below 105%. Even the genuine expansion is concentrated. --- Each type tells a different story. Indexation is lock-in, customers absorbing the increase because leaving costs more than paying it. Usage could be love if it reflects adoption, or circumstance if the customer's headcount simply grew. Cross-sell is the clearest signal of product value. Strip out the price increase and the single large account, and genuine expansion gets you to about 95% net retention. Healthy, but a long way from 115%. --- You walked into the board meeting with one green number. You're leaving with a renewal rate worse than the retention rate, expansion built mostly on pricing and circumstance, and a headline that depends on lock-in more than love. That 115% was never one number. The dashboard said compounding. The data says it depends. ---