The 60-second time-to-value benchmark

The benchmark moved because the product now runs before the account does

Wes Bush spent years teaching product-led teams to obsess over time-to-value, and the working number was 10 minutes: get a user to a real “aha” inside the first session or lose them. This year he reset it publicly, and the new number is roughly one minute. That is not a tightening of the same target. It is a different target with a different architecture underneath it.

The old 10-minute window assumed a sequence: sign up, configure, connect data, then experience value. Every step in that sequence was friction the user tolerated because the payoff was real and there was no faster path. AI-native products broke the sequence. A user can now describe what they want in plain language and watch the product produce something useful before an account exists. The value is no longer at the end of the setup funnel. It is the first thing the user touches.

That inversion is why the benchmark collapsed by an order of magnitude. When the product can generate output from a prompt, the cost of “trying it” drops to the cost of typing a sentence. Any onboarding step that sits between the user and that first output is now measured against a competitor who removed the step. Sixty seconds is not an aspiration. It is what the market already ships.

Three architectural changes make 60 seconds possible

The first change is that the product runs on a shared model, not the user’s data. PLG 1.0 value was locked behind integration: connect your CRM, import your contacts, sync your repo. The product was empty until you filled it. AI-native products arrive full. The model already knows how to write the email, summarize the document, generate the query. The user’s data makes the output better, not possible. That single shift moves the first “aha” from post-integration to pre-signup.

The second change is that input is now conversational, and conversation carries no setup cost. A form has fields, and fields have to be understood, validated, and filled in the right order. A prompt has none of that. The user states intent in the language they already use, and the product infers the rest. The interface stopped being a configuration surface and became an intent surface. You do not teach the user your schema. You read theirs.

The third change is that inference is cheap enough to spend on strangers. Serving a first result to an anonymous visitor used to be uneconomical for most products, so value was gated behind an account as a cost-control measure disguised as a growth step. That math flipped. Running a model against one prompt from an unauthenticated user costs a fraction of a cent, and the conversion lift from letting them see the output first pays for it many times over. The gate did not move for design reasons. It moved because the unit economics stopped defending it.

None of this works without trust in what the product returns. A pre-signup output that is confidently wrong is worse than no output, because it fails the user at the exact moment they were deciding whether to believe you. The 60-second benchmark is a trust benchmark before it is a speed benchmark. The teams hitting it are the ones treating the first result as a promise rather than a demo.

The interactive walkthrough is dead weight now

The dominant onboarding pattern of the last decade was the guided tour: tooltips, coach marks, a checklist of “complete your profile” tasks, a product tour that walked the user through features they had not asked to see. That pattern existed to bridge the gap between signup and value. It was scaffolding around a slow first experience.

When value lands in the first 60 seconds, the scaffolding becomes friction. A walkthrough that interrupts a user who is already getting a useful result is not helping. It is delaying the thing they came for. The best onboarding for an AI-native product is often no onboarding at all: the user prompts, the product responds, and the “aha” happens without a single tooltip.

This is uncomfortable for product teams because interactive walkthroughs were a measurable, ownable surface. You could A/B test a checklist. You could report activation lift from a redesigned tour. Killing that surface means giving up a familiar lever. But a lever that optimizes a step the product no longer needs is optimizing waste. The activation work moves upstream, into the quality of the first output and the clarity of the intent surface that produces it. That is harder to instrument and far more valuable.

The counterargument is that complex products still need teaching, and a 60-second output does not make a user proficient. True. A first result is not mastery, and products with real depth still owe the user a path from first value to fluent use. But that path starts after the user is convinced, not before. The mistake is front-loading the teaching into the moment the user is still deciding whether to care. The first minute has to earn the second one.

The teams that will get left behind are the ones defending the funnel

The teams at risk are not the ones with weak AI features. They are the ones whose growth model still assumes the old sequence. If your activation dashboard measures signup-to-value and treats the account as step one, you are instrumenting a funnel that starts after your competitor already delivered value. You will optimize the wrong minutes and report progress while losing the users who never signed up because a rival showed them the result first. This is the same gap between adopting PLG and actually tracking it that shows up across the activation data, where the signals that predict conversion sit earlier than most teams look.

The tell is where the gate sits. Ask a team to draw their funnel and watch where the signup wall goes. If value is behind the wall, they are running PLG 1.0 in a market that reset the benchmark. The real predictive signals now live in the sequence of features a user adopts first, not in the account they create. Moving the wall is not a UX tweak. It touches billing, abuse prevention, model-serving cost, and the definition of activation itself. Teams that treat it as a landing-page experiment will keep losing to teams that rebuilt the architecture.

The 60-second number is not the point. The point is that the first useful thing a product does now happens before the relationship is formalized, and every growth assumption built on the reverse order is quietly obsolete. The benchmark moved because the product moved in front of the funnel, and the teams that follow it there will own the users the funnel never sees.