A/B Testing for Service Businesses: Plumbers, Lawyers, Dentists, and Everyone Else
Most A/B testing advice is written for online shops. If your website's job is to get someone to call you, fill out a form, or book an appointment,...
A · Control
Sample experiment
A/B testing · built for small teams
Everything you need to run experiments.
SplitPea helps small teams run focused A/B tests without an infrastructure overhaul. Make one change, measure what visitors do, and get a clear recommendation when the evidence is strong enough.
Conversion rate3.1%
5,142 visits160 signups
SplitPea
Built for small teams
A/B testing without the enterprise overhead
A/B testing built for small businesses.
SplitPea gives small teams precision split-testing with no data scientist and no code. Every decision is backed by a number you can trust.
Sample result+51%
10,260 visits96% confidence
Install
1 snippet
Client size
Under 8KB
Winner threshold
90%+
Starter plan
$0
01 · What you get
Six features.
Six features.
That's the whole product.
Experiment builder
Name the test, write down what you think will happen, pick the page, and choose what counts as a win. The builder keeps those decisions in a few focused steps.
Experiment briefReady to draft
HypothesisShorter headline → more signups
GoalClick “Start free”
Visual editor
Click the headline, button, image, or block you want to change. Type the new version. No code and no dev ticket.
Traffic splitting
Half your visitors see the original, half see the variant. The same person keeps seeing the same version.
Goal tracking
Pick the action that matters: a signup, click, form submission, booking, or page visit.
Plain results
Visitors, conversions, lift, and confidence in normal language. If the result is unclear, SplitPea says so.
Decision log
Every test ends with ship A, ship B, or inconclusive, plus the date, the person, and the reason.
02 · The product
An experiment, read at a glance.
One screenOne recommendation
Recommendation96% confidence
Next action
Ship variant B
One screen tells you which version is ahead, by how much, and whether the result is ready to act on.
Visitors
10,260
Across both variants
Control signups
160
3.1% CVR
Variant signups
241
4.7% CVR
Relative lift
+51%
Variant B
Confidence
96%
Clear result
Decision
Ship B
Logged for later
03 · How it works
Four steps.
Four steps.
One simple workflow.
Most A/B testing tools assume you have a growth team. SplitPea keeps the workflow smaller: install, test, measure, decide.
01
Install
Drop one snippet on the site you already have. It works with static HTML, Squarespace, Webflow, WordPress, Shopify, and more.
<script src="sp.js">
02
Compose
Write a hypothesis, open the visual editor, and make one useful change. Your live site stays untouched until the test starts.
ControlCurrent page
VariantNew headline
03
Measure
Watch traffic split and confidence build. SplitPea tells you when there is enough data, and when there is not.
Confidence72%
04
Decide
Ship the winner, hold the original, or record an inconclusive result. The decision stays in the log.
→ Ship variant B
04 · Install
One snippet. Install it once.
Paste the snippet into the head of your site. It is under 8KB, loads asynchronously, and works with the stack you already have. Check it on your own site before launching a test.
<!-- SplitPea: paste once, leave it -->
<script src="https://cdn.splitpea.co/sp.js"
data-key="YOUR-SITE-KEY" async></script>
<script src="https://cdn.splitpea.co/sp.js"
data-key="YOUR-SITE-KEY" async></script>
05 · Experiments and insights
Sample decision log.
A sample log showing how SplitPea records what was tested, what happened, and what you decided.
| Date | Who decided | What was tested | Verdict | Lift |
|---|---|---|---|---|
| 03 May | MaraLinen & Co · Founder | Plain pricing against feature-list pricing on /pricing. | Shipped B | +51% |
| 28 Apr | JonasOutpost Studio · Designer | Three subhead variants on the home hero. | Held | — |
| 21 Apr | PriyaNorthbeam · Marketing lead | Cut the lead form from seven fields to four. | Shipped | +18% |
| 14 Apr | BenFolio · Solo founder | Trust badges above vs. below the CTA. | No change | 0.4% |
| 07 Apr | ChiaraMossroom · Growth | “Get started” vs. “Run my first test”. | Shipped | +9% |
Sample records for illustration. Your own experiment history exports as CSV.
06 · Subscriptions
Three plans.
Three plans.
Plain prices.
Start free, upgrade when testing becomes a habit. Every plan includes the visual editor, confidence scoring, goal tracking, and decision history.
Plan 01 — Starter
$0
free
Try it on one site.
✓1 website
✓1 active experiment
✓500 visitors per month
✓Visual editor and goal tracking
✓Confidence scoring and decision history
Recommended
Choose Grow
Plan 02 — Grow
$29
/mo
For regular testing.
✓3 websites
✓5 active experiments
✓5,000 visitors per month
✓Visual editor and goal tracking
✓Confidence scoring and decision history
Plan 03 — Agency
$49
/mo
For client sites.
✓Unlimited websites
✓Unlimited active experiments
✓50,000 visitors per month
✓Visual editor and goal tracking
✓Confidence scoring and decision history
07 · Footnotes
FAQs.
Will the snippet slow my site down?
The snippet is under 8KB and loads asynchronously to keep the added work small. As with any script, check your own site before launching a test.
Does it cause flicker?
SplitPea uses an anti-flicker hook to reduce visible page changes. The result still depends on the page, connection, and size of the variant, so preview the test before launch.
What happens if there is no winner?
That is a real result. SplitPea marks the test inconclusive so you can stop waiting and test something more meaningful.
Can I export my decisions?
Yes. Experiments, variants, results, and decisions export as CSV. Your evidence is yours.
From the blog
Read the blog
Experiment notes for modern makers.
Getting Started
A/B
Getting Started
A/B
How Much Traffic Do You Need for A/B Testing?
The honest answer to the question every small business asks before running their first test. Real numbers, not "it depends."
Getting Started
A/B
A/B Testing for Static Sites (Without a Server or a $200/Month Tool)
Most A/B testing guides assume you've got a server rendering your pages. If your site is plain HTML, built with Hugo or Eleventy, or hosted on...
Pick one thing you want to know
Start with one clear question.
Write a hypothesis, draft one variant, and check the preview before starting the test. Once the snippet is installed, creating the experiment should take only a few minutes.