Hire an A/B test developer
Hire an A/B test developer who ships in days, not sprints
A senior experiment developer, by the test or the month: 24–48 hour builds, QA’d and tracking-verified, delivered white-label when the client is yours.
You are making a hiring decision, so read this like a job spec. The terms are written down, the turnaround is a standing commitment, and the trial is a paid pilot: one real brief, one fixed fee agreed before work starts, one build with the QA sheet attached, then you decide. You keep the strategy, the analytics and the code. We are the hands.
- 24–48h from brief to launch-ready
- Paid pilot: one brief, fixed fee, then decide
- White-label, under your own NDA
Hired into your stack and your workflow
The problem
The four ways teams staff test development
When the backlog outgrows the build capacity, the same four answers come up, and each one charges for something you did not ask for.
The in-house hire
Around £72,000 a year all-in at UK rates, months of recruiting, a six-month ramp, and a single point of failure who can resign the week before peak, for a workload that spikes before peak season and dips inside it.
The freelancer
The lowest price per test, no cover when they are ill, no analyst behind them, and no written QA standard. The discount in a cheap build is usually the QA that was never done.
The borrowed product engineer
Free, in the sense that the salary is already being paid. But their queue belongs to the roadmap, and an experiment ticket loses to a feature ticket every sprint, so the two-day build arrives next quarter or never.
The retainer you did not ask for
Ask some agencies for a pair of hands and a strategy programme comes back. You wanted a developer; the quote wants to own your roadmap.
We sell the pair of hands, straight: senior developers who build experiments all day, hired per test or per month, QA sheet attached, code yours, and no strategy pitch stapled to the invoice.
Hire from a team 1640+ shipped tests deep.
A developer, not a programme
Hands for your programme. Not a pitch for ours.
If you want the whole loop run for you, that is our A/B test development service; on Shopify, the Shopify CRO programme. If you do not know what to test yet, start with the free CRO audit. You already know what to test. What you need is someone to build it.
What we build
What a delivered test looks like
The code you receive
- Written by hand and loaded ahead of first paint, so nothing flashes before the variation renders: on our last standing programme the median added paint delay was 12ms
- A fallback-to-control wrapper on every variant, so anything unexpected degrades to the original page instead of a broken one
- Readable, commented and portable, because it is your code: into another tool, another developer’s hands, or your theme
The QA that precedes handover
- The 20-check pre-launch checklist we publish as a free template, run in full on every build: brief, rendering, tracking, and the first hour of live traffic
- The full browser and device matrix: Chrome, Safari, Firefox and Edge, on iOS, Android and tablet breakpoints
- One rule governs it: whoever wrote the variation does not sign it off
Tracking proved, not assumed
- Every goal, click and revenue event verified against real payloads before a visitor sees the test, because a beautiful variant with silent tracking returns four weeks of nothing
- Traffic assignment watched through the first hour and checked for sample ratio mismatch, the failure mode our free SRM checker exists for
- Events wired for the readout you actually run: GA4, Mixpanel or Amplitude
White-label and NDA terms, in writing
- Your NDA signed before the first brief, and every ticket, thread and review lives in your tools under your process rather than ours
- Everything you hand onward carries your branding or none: the staging link, the defect log, the signed QA sheet
- No Optyv fingerprints in code comments or commit history, the same standard our standing white-label retainer is built on
The commercial terms, also in writing
- A fixed fee, quoted and agreed before work starts, never a meter running on clarifications
- The variation code is yours mid-engagement and at exit, so switching costs stay a line item instead of becoming the relationship
- Per-test or monthly capacity, and either one stops at the end of any month
The difference
Renting the capability vs. hiring it
| An Optyv test developer | An in-house test developer | |
|---|---|---|
| Time to first build | The next brief you send, at 24–48h standard turnaround | Months of recruiting, then a six-month ramp |
| Cost shape | A fee per test, or monthly capacity, used or paused | Around £72,000 a year all-in, busy or not |
| QA standard | A published 20-check matrix, signed by someone who did not write the code | Whatever one person thinks to check, until you write it down |
| Cover and continuity | A team standard that survives any one person’s holiday | A single point of failure who can resign before peak |
| Exit terms | The code is yours; stop at the end of any month | Notice periods, and a handover you have to chase |
Why us
Three ways to hire a test developer
| Engagement | What you get | How it is bought |
|---|---|---|
| Paid pilot | One real brief from your backlog, built and QA’d to the same standard as a retainer build, with the completed QA sheet attached. You judge the scope response, the code and the communication, then decide. | A fixed fee, quoted and agreed before work starts. No commitment past it, and no retainer pitched on the readout. |
| Single test builds | One brief in, one live experiment out, watched through its first hours. The shape that fits when you have a strategist already and the backlog arrives in bursts. | A fixed price per test, quoted before the build starts. Simple copy and layout changes cost least; multi-page, server-side and hostile-theme builds cost most. |
| Monthly capacity | A standing pair of hands for a steady backlog, working at a reserved cadence your product roadmap has no claim on. The workhorse shape for agencies and in-house teams alike. | A monthly fee, scaled to your cadence. Stop at the end of any month. |
How it runs
From job spec to standing capacity
The same route for an agency white-labelling delivery and an in-house team buying throughput: prove it on one paid build, then choose a shape. Nothing is sold before the pilot has been judged.
- 1
Send the job spec
A brief, a Figma file, or the job advert you were about to post. We review it same day, flag the edge cases and the data-layer questions up front, and answer with a fixed fee and a delivery date. If the brief is not buildable as sent, we say so before quoting rather than after invoicing.
- An answer the day your spec lands
- Fixed fee quoted before work starts
- Unbuildable briefs named before invoicing
- 2
The pilot build
One test, built for real: hand-written variation code in your platform, every goal and revenue event wired and verified, the full 20-check QA matrix run and signed by someone who did not write the code. Standard builds land in 24–48 hours; complex ones are scoped honestly at 3–5 business days.
- 24–48h standard, 3–5 days complex
- Full QA matrix, signed off independently
- Tracking verified against real payloads
- 3
Judge it, then choose a shape
The pilot hands you everything you need to decide: the code, the QA sheet, the communication trail. Stay per-test, reserve monthly capacity, or take the build and leave, it is your code either way. Agencies sign the white-label terms here: NDA, unbranded delivery, our name nowhere your client looks.
- Per-test, monthly capacity, or walk away
- NDA and white-label terms signed for agencies
- The pilot code is yours whatever you choose
- 4
A standing pair of hands
Briefs in, QA’d builds out, winners implemented, and capacity reconciled monthly against what actually shipped rather than what was forecast. On our standing white-label retainer that rhythm holds a 2.1-day median from approved brief to client-ready link, with client QA bounces down from 31% of builds to 2%. Your strategy stays yours; the build is never the bottleneck again.
- Median 2.1-day turnaround on our standing retainer
- Monthly reconciliation against shipped builds
- Stop at the end of any month
What hiring us includes that a hire does not.
NDA before the brief
Signed before your first brief lands, so nothing is discussed unprotected.
Cover, not a person
A team behind every build, so a holiday or a resignation changes nothing.
Signed off by someone else
Whoever writes a variation never QAs it. Self-review misses what that rule catches.
Code you can walk away with
Readable and commented, yours to take mid-engagement or at exit.
Statistics we will not flatter
SRM checked in the first hour, and no guaranteed win rates, ever.
Capacity that flexes
A heavy month before peak, a quiet one inside it, and no salary to renegotiate.
They handed us their whole test playbook when we hired in-house. Who does that? Agencies that aren’t scared.
GGeorgia TanNorthbeam Outfitters
Technically reviewed by Zahidul IslamCTO & Experimentation LeadThe answers to your questions.
Two shapes, both quoted as a fixed fee before work starts: per test, or standing monthly capacity scaled to your cadence. What moves a per-test fee is how much of the page moves, how hostile the theme is and whether the work is server-side, so simple copy and layout variations cost least and multi-page or server-side builds most. What moves a monthly fee is the cadence you reserve. Neither shape asks you for a lock-in, and the monthly one stops at the end of any month. Send one real brief and the quote comes back with a delivery date attached.
You send one real brief, we quote a fixed fee and agree it before work starts, and you get a production-grade build with the completed QA sheet attached. Then you decide: keep sending briefs per test, reserve monthly capacity, or walk away with the code, which is yours either way. No retainer is pitched on the pilot readout, ever.
On the next brief you send, because there is no onboarding project in front of it: no discovery phase to buy, no tooling to migrate, no ramp to fund. Compare that with the hire this replaces, where the recruiting, the notice period and the ramp all happen before a single variation ships. On the white-label retainer we have run for ten months, the median from approved brief to client-ready link now sits at 2.1 days, against the 16 business days that agency’s freelancers took before us.
You do, mid-engagement and at exit. Every variation ships as readable, commented code you can move into another tool, hand to an in-house developer or rebuild into your theme, and nothing about leaving us changes that. Code you cannot take with you turns a supplier into a switching cost, and we refuse to be that vendor.
No, and the mechanics are written into the engagement rather than promised on a call. We sign your NDA before the first brief. Our name appears nowhere in code comments or commit history. Questions a brief leaves open come back to you, never to your client, because the arrangement gives us no access to end-client stakeholders at all, by design. White-label is the standard agency arrangement here, not a paid upgrade.
No: you are hiring into your stack, not migrating onto ours. We build natively in Optimizely, VWO, Convert.com, AB Tasty and Adobe Target, across Shopify Plus, Magento / Adobe Commerce, WooCommerce and headless storefronts, so the platform question is usually settled in one sentence. Process matters as much as platform for a hire: briefs arrive in your ticket format, builds land on your board, and we hold each build to your QA tier rather than asking you to adopt ours. One current engagement runs three different testing platforms at once across a single client roster.
You pay for the build, not the result, and a cleanly built loser is a real answer: it killed an idea with data before the idea was rolled out to every visitor. Across our archive roughly one test in four or five wins, and anyone quoting you a higher hit rate is describing their statistics, not their developers. What we guarantee is the part a developer can guarantee: the variation renders without flicker, the tracking fires, and if anything breaks, the visitor gets your original page instead.
Same developers, different amount of us. Hiring here buys a pair of expert hands on the programme you already run, priced per test or per month, with the strategy and the verdicts staying yours. Our A/B test development service is the wider execution offer, and if you do not yet know what to test, start with the free CRO audit instead of either. If you want the whole loop run for you, strategy included, that is the development service, not this one.
Hired hands, measured in production
Skip the recruiting cycle
Send the brief you would have handed a new hire, or just the job spec you were about to post. We reply to all inquiries within one working day with a fixed pilot fee and a delivery date, and if the honest answer is that you should hire in-house instead, we say that too.
Brief your first test build ↗