A/B Testing for Make (Integromat) Automations
Make (formerly Integromat) lets you build sophisticated visual automations that connect hundreds of apps without code. Its routers, filters, and modules make it ideal not just for moving data, but for running controlled experiments on the automated journeys your customers travel. This guide covers how to split scenarios, test variant paths, and turn your Make automations into a measurable optimization program.
Where A/B Testing Fits in a Make Scenario
Make scenarios are built from modules connected by routes, and the Router module is the natural home for an A/B test. By adding a router with a randomizing filter, you can send each bundle down one of two (or more) paths and compare the downstream results. This works for any automated decision: which email template to send, which delay to apply, which offer to surface.
Because Make processes data in discrete bundles with full execution history, you get a built-in audit trail. Every scenario run is logged, so you can reconstruct exactly which path each record took and what happened next — provided you store the variant label somewhere durable.
- ● Use a Router with a randomizing filter to split records 50/50 between two paths
- ● Test two onboarding email templates and compare reply or activation rates
- ● A/B test offer timing: send a discount immediately vs. after a 24-hour delay
- ● Split a lead-scoring scenario and compare which threshold produces more qualified handoffs
Setting Up a Clean Split
To split traffic, add a module early that generates or reads a stable random value for each record — Make's built-in functions or a data store can hold the assignment so the same record always follows the same path. Then configure router filters so each branch only accepts its assigned bucket. Always persist the variant label to a Make Data Store, a spreadsheet, or your CRM before the paths diverge.
The discipline that separates a real test from guesswork is paired measurement: record the assignment at the split, then record the outcome when the conversion fires, both keyed to the same record ID. Make Data Stores are perfect for this because they persist across scenario runs.
- ● Store variant assignments in a Make Data Store keyed by contact ID
- ● Use filter conditions on each router branch to enforce a clean split
- ● Write the conversion outcome back to the same Data Store record for analysis
- ● Define your success metric and minimum sample size before turning the scenario on
Pairing Make with On-Page Testing
Make shines at automating what happens after a conversion, but the conversion itself usually depends on the page — the landing page headline, the form, the pricing layout. Make cannot test those on-page elements, so the strongest programs run the page experiment in a dedicated tool and let Make react to the result.
Connect a tool like abTestBot to a Make webhook so each conversion arrives tagged with its variant. Your scenario can then route follow-ups, update records, and trigger notifications while staying fully aware of which page version won the visitor over — closing the loop between page test and automation.
- ● Trigger a Make scenario from an abTestBot conversion webhook, carrying the variant label
- ● Route post-conversion emails differently based on which page variant converted the user
- ● Sync the winning variant into your CRM custom fields automatically
Avoiding Contaminated Results
Resist the urge to tweak a live scenario. Changing a module, a filter, or a template mid-test breaks the link between cause and effect and forces you to restart. Treat a running experiment as frozen, and clone the scenario if you need to keep iterating on something else in parallel.
Also watch for sample imbalance. Make scenarios can fail or skip bundles for unrelated reasons (rate limits, app errors), which can quietly skew your split. Monitor the per-branch counts and investigate any large imbalance before you trust the result.
- ● Clone scenarios rather than editing a live test, so the experiment runs unchanged
- ● Monitor execution logs for errors that could bias one branch over the other
- ● Run for two full business cycles and use proper significance testing before deciding
Ready to Start Testing?
abTestBot handles the on-page split and feeds variant-tagged conversions into your Make scenarios automatically. Start optimizing the pages behind your automations today.
Get Started Free →Related Guides
A/B Testing with n8n: Automate Your Experiment Workflows
Use n8n to automate A/B testing — split traffic, trigger experiments, sync results to your data warehouse, and test automated email and onboarding flows.
A/B Testing with Zapier: Test Your Automated Funnels
Discover how to A/B test with Zapier — use Paths to split automated workflows, test email and SMS sequences, and measure which automation converts best.
A/B Testing for Email Marketing: Subject Lines to Send Times
Optimize every email with A/B testing. Learn how to test subject lines, preview text, send times, layouts, and CTAs for higher opens and clicks.