Grow
Conversion Rate Optimization & Experimentation
Get more sales from the traffic you already pay for. A/B testing for Shopify and DTC stores, and experimentation programs for larger teams.
Readout
+111%
Subscription revenue in four months, Fresh Peaches
Shawn has run testing programs at LegalZoom, Samsung, MeUndies, CBS Interactive and Fresh Peaches.
Where stores lose sales
“We get traffic, but nobody buys.”
→ Conversion rate
“People add to cart, then disappear.”
→ Cart & checkout abandonment
“Our average order is too small.”
→ Average order value
“Product pages get views but no action.”
→ Add-to-cart rate
“Customers buy once and never come back.”
→ Repeat & subscription rate
What a testing program includes
Testing focused on the pages that drive revenue: product pages, the cart, checkout, and how offers and subscriptions are presented.
| Includes | |
|---|---|
| Research | Funnel and analytics audit, UX review, heatmaps and session recordings, customer surveys, and mining reviews and support tickets |
| Roadmap | A prioritized list of tests, scored by expected impact, confidence and effort |
| Build & QA | Variants designed and built in your theme or testing tool, checked across devices |
| Readouts | Results with confidence intervals, what we learned, and what to test next |
Examples of what we test
Does leading the product page with how-to-use photos beat leading with ingredients?
Does defaulting to subscribe-and-save raise subscriptions without hurting first orders?
Do bundle offers raise AOV more in the cart or on the product page?
Does showing shipping cost on the product page reduce checkout abandonment?
Do reviews near the top of the page beat reviews below the description?
The first few months
| When | What happens |
|---|---|
| Weeks 1–2 | Research and quick fixes at the same time: analytics audit, session recordings set up, obvious problems fixed |
| Each month | New tests, winners shipped, and a readout of what we learned |
| Each quarter | The roadmap reviewed against revenue goals |
How we call a winner
It's common to see a "+38% lift" reported from a test that never had enough traffic to prove it. We size tests so the result means something.
One primary metric, chosen up frontUsually revenue per visitor rather than conversion rate alone. A test that raises conversions by discounting everything isn't a win.
Guardrail metricsWe track AOV, refund rate and subscription churn so a winning variant can't quietly cost you money.
Sample size and runtime set before launchNo checking early and stopping when the chart looks good.
Clear about traffic limitsEach test is sized from your conversion rate and the smallest lift worth detecting. If traffic can't support a clean test, we change the approach (bigger changes, longer runs or research-led fixes) and keep the same standard of proof.
Losing tests countA loss tells you what customers don't want. Every result gets written up.
Before you pay for a test, we'll tell you whether it's worth running.
Experimentation programs for larger teams
For companies that want testing to be part of how they work. Shawn has built experimentation programs from scratch inside large companies and trained the analysts and product managers who ran them.
Program setup
Testing tool selection, a shared backlog, and one way of prioritizing that the whole team uses.
Test governance
Hypotheses written down before launch, primary metrics and guardrails, sample-size planning and standard readouts.
Team training
Analysts and product managers trained to design, run and read tests on their own.
Forecasting
Conversion forecasts accurate enough for finance to plan with, so testing connects to the business plan.
Conversion work, with the numbers
- OwnedFresh Peaches
+111% subscription revenue in four months from a conversion-optimized subscription widget, plus +36% onsite conversion year over year.
Read the case → - In-houseSamsung
A 3% increase in conversion rate for mobile phones, from a rebuilt purchase flow proven in an A/B test.
Read the case → - In-houseCBS Interactive
Built the A/B testing program from scratch, trained the analysts and PMs, and built a conversion forecast accurate to 3–8% MAPE.
Read the case →
Why not a testing tool, a hire or your build agency?
| Testing tool alone | In-house CRO hire | Your build agency | Forepoint | |
|---|---|---|---|---|
| Knows what to test | No, it's software | Eventually | Guesses | A research-led roadmap |
| Statistical rigor | Depends who runs it | Depends who you hire | Rarely | A former experimentation lead |
| Time to first test | Weeks of setup | 3–6 months to hire | N/A | Weeks |
| Builds the variants | No | Needs a developer | Yes | Yes |
Your build agency launches the store. We work on getting more sales out of it.
How engagements run
Start with a CRO auditA ranked list of what's costing you conversions, with the reasoning and expected impact for each item.
Then a monthly testing programResearch, tests, readouts and shipped winners on a regular schedule.
If the problem isn't on your website…
Revenue Optimization →Common questions
If yours isn't here, just ask.
How much traffic do I need for A/B testing?
It depends on your conversion rate and the size of lift you care about, and we work it out before any test runs. As a rough guide, a clean result often needs a few hundred conversions per variant. Below that, we test bigger changes, run tests longer, or ship research-led fixes and measure them carefully, and we're clear about what the data can and can't support.
Which A/B testing tools do you use?
Whatever suits your store: Shopify testing apps, VWO, Optimizely or others. We've run programs on enterprise tools and lightweight ones. The discipline matters more than the tool.
Who builds the test variants?
We do. We design and build variants in your theme or testing tool and check them across devices before launch.
How long until we see results?
Quick fixes can go live in the first two weeks. A properly sized test usually runs two to four weeks, so most programs have a first readout within the first month or two.
What happens if a test loses?
You learn something. A losing test tells you what customers don't respond to, which makes the next idea better. We write up every result.
Do you work with larger companies, not just Shopify stores?
Yes. For larger teams we set up and run experimentation programs, including the tooling, backlog, governance and training, so testing continues after we step back.
Do you work on Shopify Plus or standard Shopify?
Both. Some tests, like checkout changes, need Shopify Plus. We'll tell you which ideas your plan supports.
Do you test pricing and bundles?
We test how prices and bundles are presented: placement, framing and defaults. Setting the price levels themselves is part of Revenue Optimization.
Next step
Tell us what you'd like to test.
Answer a few questions about your business, then pick a time for a 45-minute intro call with Shawn.