The essentials
- The answer: no A/B test in the strict sense (two versions served in parallel) — Google offers none, and a listing has only one version.
- What does work: the sequential test (period A then period B on the same listing) and the parallel test between comparable listings in a network.
- What can be tested: the cover photo, the type and frequency of posts, secondary categories, the description, the action button, the offer.
- What can't be tested: the primary category and the name, where changes trigger re-verification and lost rankings.
A/B testing means exposing two versions of an element to comparable audiences and measuring which performs better. On a website or in Google Ads, it is routine. On a Google Business Profile listing, it isn't provided for: the listing has one version, served to everybody. You can still test, provided you adapt the method and accept less precision.
Why a true A/B test is impossible on a listing
- Google offers no variant function; every change applies to 100% of users.
- The listing's statistics are aggregated (views, calls, directions, clicks) with no segmentation by version or by user.
- The data comes with a delay of a few days and is rounded, which limits short tests.
- Ranking depends on distance: two users don't see the same listing at the same rank, which makes the comparison noisy. See understanding local ranking.
Method 1: the sequential test on one listing
You measure a baseline period, change a single element, measure an equivalent period, and compare. Validity depends on respecting a few rules.
| Rule | Why |
|---|---|
| One change at a time | Otherwise the effect can't be attributed |
| Periods of at least 4 weeks, aligned on the same days | To smooth weekly seasonality and the statistics delay |
| Compare conversion metrics (calls, directions, clicks, messages) and not only views | Views depend mostly on ranking, not on the element tested |
| Control the context | No Ads campaign, holiday or event falling in only one of the two periods |
| Require a gap of at least 15 to 20% | Below that it is noise, at the volumes of a small business listing |
Example: a bakery replaces its cover photo (the shop front) with a photo of its signature product for four weeks. Direction requests go from 120 to 155 at equivalent listing traffic: the 29% gap justifies keeping the new photo.
Method 2: the parallel test across a network's listings
A network of 10 locations or more can run genuine comparative tests: apply version A to half the listings and version B to the other half, over the same period. The conditions: comparable listings (the same type of area, similar volumes), a random or matched split (one city-centre listing on A for one city-centre listing on B), and consolidated tracking. It is the most reliable method; it requires a bulk publishing and reporting tool, described in the bulk publishing tools.
What is worth testing
| Element | Metric to track | Risk |
|---|---|---|
| Cover photo and profile photo | Clicks to the site, directions, calls | None |
| Frequency and type of posts (an offer, a project, an event) | Clicks on the post's button, calls | None |
| The description (angle, keywords, call to action) | Impressions on discovery queries, clicks | Poor |
| Secondary categories | Impressions on new queries (the listing's Search Console, rank tracking) | Low to moderate |
| The action button (book, get a quote, order) | Clicks on the button | None |
| Detailed services and products | Impressions, calls | None |
| The review reply template | Average rating, volume of subsequent reviews | None |
Don't test the primary category or the name: changing them can trigger a new verification, a temporary suspension and a loss of rankings that takes weeks to recover. Set them after analysing your competitors, as described in optimising your Google Business Profile.
Measuring cleanly
- Export the listing's statistics each month (calls, directions, clicks, messages, views by source) into a dated table.
- Add UTMs to the site link so you can measure in GA4 the traffic and conversions coming from the listing; see UTMs and the Google listing.
- Track rankings by grid so you can tell a conversion effect (the element tested) from a ranking effect (outside the test); see tracking your position on Google Maps.
- Log every change with its date: that journal is what makes the tests readable afterwards.
How GreenRed helps
Rather than juggling several tools, GreenRed's Google Business Profile module brings these metrics together in a single dashboard, compares them over time and tells you which actions come first. You can try it free, with no card, from the Pricing.
Frequently asked questions
Does Google offer tests on listings?
No. Google Business Profile has no variant or experiment function. The only visible experiments are the ones Google runs itself on how listings display, outside the business's control.
How long should a sequential test last?
Four weeks per period at minimum, so eight weeks for a test, to smooth weekly seasonality and the statistics update delay. For a low-volume listing (fewer than 50 enquiries a month), extend to six or eight weeks per period.
Can you test keywords in the listing name?
No. Adding keywords to the name breaks Google's rules and exposes the listing to suspension. The name must be the business name as it appears on the shop front and on official documents.
Do listing tests affect ranking?
The testable elements (photos, posts, buttons, description) act mainly on conversion. Secondary categories and services can widen the queries the listing appears on. Ranking depends mostly on the primary category, reviews, distance and consistency of information.