A Short History of Why SaaS Boards Started Obsessing Over NDR

How board reporting evolved

Pre-2015: Logo Churn & Customer Count

Board decks tracked customer count and basic logo churn. Lose 5% of your customers, report 5% churn, move on.

This told you nothing about whether the remaining 95% were expanding or contracting, a real blind spot that nobody was worried about yet.

2015–2019: Gross Revenue Retention & Broad Expansion Tracking

Top-tier VCs noticed that the most valuable SaaS companies (Box, Okta, HubSpot) shared something in common: "net negative churn."

Finance teams started tracking Gross Revenue Retention and dollar-based expansion, but usually as an internal health check, not the headline number shown to the board.

Post-COVID to today: Net Dollar Retention Commands the Room

NDR became the gold standard almost overnight, and it hasn't given it up since.

What actually triggered the shift

When COVID hit in early 2020, the old growth model broke in a specific, traceable way:

• The new-logo freeze. Corporate budgets froze worldwide. Acquiring new enterprise logos became slow, expensive, and difficult almost overnight.
• The flight to capital efficiency. Investors stopped rewarding raw ARR growth and started demanding proof a business could fund itself.
• NDR as the survival metric. Boards realized the cheapest, most predictable way to grow during a crisis was expanding the customers you already had. NDR is the metric that measures exactly that.

The valuation swing, in numbers

A few things worth pulling out of that table directly:

• Multiples roughly tripled during the COVID peak, then gave almost all of it back within two years. That's the kind of swing that makes "raw growth" a dangerous thing to optimize for in isolation.
• The premium for elite companies has consistently outpaced the median in every era, but what earned that premium changed completely: growth rate alone, then NDR and Rule of 40, then free cash flow today.
• According to McKinsey's analysis of more than 100 B2B SaaS companies, top-quartile companies by valuation multiple traded at a median 24x enterprise-value-to-revenue, compared to 5x for bottom-quartile peers. Net revenue retention was one of the metrics most correlated with which quartile a company landed in.

Why boards still obsess over it today

• Compounding value. A company with 115–120% NDR can grow revenue 15–20% a year without spending a dollar on new sales or marketing.
• The clearest product-market-fit signal available. High NDR means customers are adopting deeper, upgrading tiers, and adding seats — proof people are getting enough value to pay more, not just enough to stay.

Pricing model

Per-seat pricing has been the default in SaaS for about twenty years, and it's already starting to move. Two of the bigger names in customer support software, Intercom and Zendesk, both shifted their AI products to outcome-based pricing in 2026, Intercom's Fin charges per successful resolution, and Zendesk rolled out its own version of the same idea not long after. (As of this writing, Intercom has rebranded its parent company as "Fin" and Salesforce has agreed to acquire it; the pricing model described here is unchanged, but the corporate name may not be current by the time you're reading this.) That part isn't speculation, it's already happened. The logic behind it makes sense too: if an AI agent can handle the workload of several people, charging by seat count doesn't really track value for either the vendor or the buyer anymore.

What that might mean for NDR

If pricing keeps moving away from seats and toward usage, resolutions, or outcomes, then NDR by itself might not be enough to tell the whole story going forward. You'd probably still want it, since it still captures whether you're keeping and growing revenue from existing customers, but it may end up sitting next to some newer signals, things like time-to-value, cost-to-serve, or how efficiently a company's AI is actually getting used. I don't think NDR goes away here, but I could see it becoming one of several numbers a board looks at rather than the one number that gets all the attention.

Margin pressure

Traditional SaaS got used to gross margins north of 80%, mostly because actually running the software was cheap once it was built. Running the AI behind a lot of these products isn't cheap in the same way. There's real compute cost involved, and that's already starting to show up in how investors think about the difference between a company that built AI into its core product versus one that bolted a chatbot onto something older.

I'd hold all of this a little loosely. The pricing shift is real and already happening; the rest is a reasonable read on where things seem to be heading, not something I'd want to present as settled.