What does a typical Tyrell Lab engagement look like?

Most engagements run six months and follow a Diagnose → Design → Hand-off method. The first 30 days are spent running the 47-point Experiment Readiness Audit and instrumenting the activation funnel; months two through four are spent shipping validated experiments; the final phase is a hand-off playbook and a 30-day shadow period where your team runs experiments with a partner on call.

Who actually designs the experiments?

Every experiment is designed by a Tyrell Lab partner with 8+ years of experimentation experience, typically former leads from companies like Booking.com or Stripe. We do not staff engagements with junior analysts — the person you meet on the diagnostic call is the person doing the work.

What happens after the engagement ends?

You keep the hand-off playbook, the hypothesis backlog, the experiment templates, and the statistical-rigour checklist. Seventy-one percent of clients expand scope within 90 days, but if you choose to run experiments without us you will have a documented program, not a dependency.

How does the ROI guarantee work?

If Tyrell Lab does not ship at least 8 validated experiments in the first quarter of the engagement, the engagement fee for that quarter is refunded in full. The guarantee is written into the statement of work, not buried in marketing copy.

Tyrell Lab · Est. 2019

Where ambitious products get the experiments they deserve.

A senior-only experimentation and growth studio for SaaS and subscription brands. We design, run, and learn from conversion experiments — and we hand your team the playbook to keep going without us.

Book a Free Experiment Diagnostic → Read the method

A 30-minute diagnostic with a Tyrell Lab partner. No slides, no sales call.

01
412 winning

Winning experiments shipped since 2019, drawn from 1,180 total runs for SaaS clients.
02
+32% activation lift

Average activation lift across shipped experiments within the first 90 days of engagement (median client: +27%).
03
92% retention

Client retention rate year-over-year; 71% of clients expand scope within 90 days of engagement.
04
4.9 /5 Clutch

Average rating across 63 verified client reviews on Clutch.

Chapter I · A note from a partner

Your program is owned by a partner, not handed to a junior.

Tyrell Lab was founded in 2019 by former Booking.com and Stripe experimentation leads. Our model is deliberately small: fourteen senior strategists, data scientists, and product designers, fully remote across nine countries. Every experiment is designed by a partner with eight or more years in experimentation, never delegated.

We don't sell headcount. When you hire Tyrell Lab, you get a small senior team that runs alongside your product, growth, and data org for a defined six-month engagement — and then leaves you with a measured baseline, a prioritized hypothesis backlog, and a sequencing plan your in-house team can run without us. The senior-only model is the point, not a marketing line.

— A Tyrell Lab partner

Chapter II · The method

Diagnose, Design, Hand-off — three chapters of one engagement.

Every Tyrell Lab engagement moves through the same three chapters. Nothing is open-ended. The hand-off is the deliverable.

I.
Diagnose

We start with the 47-point Experiment Readiness Audit — a rigorous diagnostic of your onboarding telemetry, activation definition, hypothesis backlog, test cadence, and statistical rigor. You receive a written scorecard in week two, before any test is queued.

Duration · Weeks 1–4 · Deliverable · Readiness scorecard and 90-day experiment roadmap
II.
Design

A senior Tyrell Lab partner runs experiments alongside your team — designing variants, instrumenting telemetry, and reviewing results with sequential testing rather than peeking at p-values. We commit to at least eight validated experiments in the first quarter, or the engagement fee is refunded in full.

Duration · Months 2–5 · Deliverable · 8+ validated experiments with documented learnings
III.
Hand-off

The final month is dedicated to transfer: a written playbook, paired shadowing of your in-house team running their own tests, and a graduated cadence that ends with you operating the program independently. We measure success by what your team can run without us in month seven.

Duration · Month 6 · Deliverable · Internal playbook, runbooks, and capability transfer

Chapter III · The diagnostic

The 47-point Experiment Readiness Audit.

The entry artifact of every engagement. A diagnostic, not a sales call — refined across 180+ SaaS onboarding engagements since 2019. Nine representative categories below.

Onboarding telemetry

Event coverage, identity resolution, and data freshness across the first-session funnel.

Activation definition

Whether your activation event is a behavioral milestone, a vanity event, or somewhere ambiguous in between.

Hypothesis backlog

Depth, prioritization, and evidence-quality of the queue feeding your experimentation program.

Test cadence

Throughput, cycle time, and the ratio of shipped wins to total experiments run.

Statistical rigor

Sequential vs. peeking methodologies, sample ratio mismatch checks, and stopping rules.

Variant discipline

Scope, isolation, and the percentage of single-variable tests in the last two quarters.

Decision capture

Whether learnings are recorded, retrieved, and re-used — or lost between quarters.

Stakeholder alignment

How product, growth, data, and design agree on what to test and why.

Tooling stack

Fit between your experimentation platform and the questions you actually need answered.

Chapter IV · Recent experiments from the lab

A working research report, not a marketing page.

Four anonymized-but-specific experiments drawn from the last twelve months. Each row is a shipped test with a recorded hypothesis, a single primary variant, a measured lift, and a final status.

№ Hypothesis Variant Lift Status

001 A redesigned first-run checklist will outperform the current welcome modal on day-1 activation. 4-step progressive checklist vs. single welcome modal +41% Shipped

002 Surfacing integration value earlier in onboarding will raise the second-session return rate. Native integration preview at step 2 vs. post-activation +22% Shipped

003 Replacing the trial-expiry banner with a usage-based nudge will extend trial-to-paid conversion. Usage-momentum nudge vs. countdown banner +18% Shipped

004 A two-question activation survey at the moment of value will outperform a single NPS prompt. Inline value-moment survey vs. post-session NPS −3% Inconclusive

Sequential testing used throughout. Negative results are recorded in the same ledger as wins.

See the full proof ledger →