Most agency proposals end with a list of deliverables. Logos, screens, components, code, ad campaigns. The contract says we'll produce them, the client signs, and three months later the artifacts arrive. Whether they did anything for the business is somebody else's problem.
Qubetix writes the metric into the SOW. Before kickoff, every engagement names a number the work has to move — onboarding time, conversion rate, signup completion, active learners — and we report the delta at handoff. Sometimes we under-shoot. We still report it.
The shape of an outcome metric
A defensible outcome metric is observable, scoped to the work, and dated. Three properties, in that order. Observable means the client's existing analytics (or ours) can read it without instrumentation theatre. Scoped means it's the slice of the funnel the engagement actually touches — not a vanity metric three steps downstream of our work. Dated means there's a measurement window: baseline → ship → 14-day delta, typically.
Three real numbers
From the case studies on this site, three engagements where the metric drove the work:
- biofideX — Target: peer-reviewed content velocity. Delivered: 5,000+ verified questions across 30+ disciplines, 10,000+ active learners across 75+ countries. The metric drove the two-reviewer content pipeline; without that constraint we'd have built a quiz app.
- Prottoy — Target: sign-up completion above 80% on mid-range Android in remote districts. Delivered: 90%+ after three auth iterations (email → SMS → phone-passcode). The metric forced us to drop SMS — the obvious choice — when it under-performed.
- Pentamart — Target: one customer order, one delivery, regardless of vendor count. Delivered: a combined-cart engine that decomposes orders into per-vendor fulfilment jobs and re-merges parcels at the destination hub. 500+ vendors, 2M+ customers.
When we miss
Two ways the studio can miss the number. The work didn't hit it (a design/engineering issue), or the metric was wrong (a discovery issue). Both happen. The contract says we'll keep working — at our cost — for two additional weeks if it's the first kind, and we'll write a short addendum if it's the second.
What we won't do is invoice for the original scope and walk away. The number is the deliverable. If we didn't move it, we owe you more work, or we owe you an honest conversation about why the number was the wrong target.
Why this shapes the engagement
Once a metric is named, three things get easier. Scope arguments dissolve — a feature is in if it serves the metric, out if it doesn't. Weekly demos have a frame — we're not showing pixels, we're showing whether the curve moved. And the handoff becomes complete: the team inherits a number they can defend at the next quarterly review, plus the codebase and brand that produced it.
It's not heroic. It's the lowest-friction way to keep design, engineering, and the business pointing the same direction for the length of a project. The work has to pay for itself; the metric is what tells you it did.
If you're sketching an engagement, the discovery call is free and ends with three written recommendations you could act on tomorrow — whether or not you work with us.