Experiment summary
- Bunzl Retail & Industry noticed that a heavily clicked category section on the homepage of its Dutch packaging shop was scattering attention across several links.
- It replaced the category blocks with customer reviews in an eight week A/B test.
- VWO reports 11% more product page visits and 9% more total sales.
- VWO supplied the testing platform and published the case. It discloses no baseline, sample size or significance, and does not say whether prices, campaigns or traffic mix changed during the test.
- The variation changed content and purpose together, so it cannot show whether fewer options, the reviews, their position or the visual treatment drove the result.
A packaging retailer removed a popular homepage feature that slowed shoppers down, then tested whether reviews offered a clearer route.
Bunzl Retail & Industry had a homepage section that looked healthy by one obvious measure. Visitors clicked it often. The section sat beneath the main banner and offered several product categories, giving customers many routes into an online shop that sells packaging and shipping supplies.
The problem appeared when the company looked beyond the click count. Its heatmap showed concentrated activity around the category blocks, while clickmaps showed attention scattered across several links. Visitors were interacting, but no single option guided enough of them further into the shop. According to the case study published by testing platform VWO, the team concluded that the choice-heavy section was creating confusion and slowing product exploration.
The distinction shaped the experiment. A busy interface can create plenty of observable behaviour without helping customers make progress. The question for Bunzl was whether the homepage needed more ways into the catalogue, or fewer competing decisions.
The clicks pointed to a bottleneck
Bunzl is an international distribution group that supplies businesses with products they need to operate. Its Dutch Retail & Industry division serves food and non-food companies with packaging, hygiene products, disposables and related logistics solutions. The experiment ran on Verzendverpakkingenshop, its online packaging store.
Interactive clicks that don’t lead anywhere useful are a symptom of a trust gap, not just a layout problem. We explored that idea more broadly in our analysis of why online retail still needs a confidence layer. Retailers running similar campaigns homepage tests can check the downstream math with our eCommerce ROI Dashboard.
The site had reasonable traffic, but visitors left the homepage quickly and the team lacked a clear explanation for the drop-off. Bunzl had formed a dedicated optimisation team in 2023 with a UX designer, developer, product owner, conversion specialist and data expert. The group used heatmaps, clickmaps, scrollmaps, funnels and session recordings to locate friction before deciding what to test.
The category area became the leading suspect because several signals lined up. Heatmaps showed intense interaction just below the banner. Clickmaps showed that attention was dispersed rather than directed. Scrollmaps confirmed that the product categories were visible above the fold, so poor visibility was not the likely explanation.
The team’s documented hypothesis was that replacing the crowded category section with customer reviews would simplify the homepage, establish credibility earlier and send more visitors toward product pages. This was a test of information priority. The categories remained available elsewhere in the shopping journey, while the homepage used its most contested space to answer a different question. Could other customers’ experiences reduce uncertainty better than another set of navigation choices?
Reviews replaced the category grid
Bunzl compared the existing homepage with a version that placed customer testimonials where the category blocks had been. VWO says the A/B test ran for eight weeks so the team could observe behaviour over a longer window rather than reacting to a short burst of traffic.
The publisher reports that the treatment increased product-page visits by 11 percent and total sales by 9 percent. Conversion specialist Hans Breeuwsma said the company implemented changes after using the behavioural evidence to form test hypotheses. The same programme later placed reviews in the shopping cart after session recordings showed visitors pausing, returning to earlier pages or leaving before confirmation.
The homepage result supports a narrow conclusion. In this test, replacing a highly interactive choice area with social proof was associated with more product exploration and sales.
The source also does not disclose the baseline sales rate, traffic allocation, sample size, confidence interval or statistical significance. VWO supplied the analytics and testing software and published the case, so the reported lifts have not been independently verified. The page does not show whether prices, campaigns, product availability or traffic mix changed during the eight-week window.
Engagement needed a second question
The useful part of Bunzl’s approach happened before the variation was built. The team did not treat the homepage’s busiest area as automatically valuable. It asked what visitors did after interacting with it, then combined several behavioural signals before changing the page.
A comparable retailer could reproduce the logic without copying the exact design. Start with a high-traffic module that receives many clicks but sends relatively few visitors into product evaluation or purchase. Keep one control. In the treatment, replace the competing choices with one form of reassurance that addresses the hesitation visible in recordings or research. Hold promotions, prices and traffic sources steady, then follow every assigned visitor through completed purchase.
The primary measure should be completed orders or contribution margin per eligible homepage visitor. Product-page visits can explain how the treatment changed behaviour, but they should remain a diagnostic measure. Before launch, the team should define the smallest commercial lift that would justify keeping the change and the level of uncertainty it will accept.
A follow-up test could separate the effects. Bunzl’s variation changed the content and purpose of the section at the same time, so it cannot tell us whether the improvement came from fewer options, the reviews themselves, their position or the visual treatment. A follow-up could separate those elements by testing a simplified category module against reviews in the same location.
The lesson is in the diagnosis
Homepage teams often protect sections that attract clicks because the activity looks like evidence of demand. Bunzl’s experiment shows why the next step deserves equal attention. A customer can interact with a page because it is useful, or because the page has made the decision harder than it needs to be.
The company used the behaviour as a prompt for a controlled comparison, then measured movement deeper into the commercial journey. The sequence is more transferable than the customer-review treatment itself. The experiment began by asking whether a popular interface element was helping customers decide, and the evidence gave Bunzl a reason to change what occupied that space.
Two other tests in this series reached a similar conclusion from different starting points. Sport Chek found that a message which had reliably lifted conversion could flip once the surrounding offer changed, and Pinterest restructured how its images pointed visitors deeper into the site rather than just changing the images themselves. In each case, the fix targeted the customer’s next step, not just the immediate click.
The Guardian’s onboarding test offers another example of changing a prominent route through a product rather than assuming more clicks mean more value.
| Experiment detail | What the reporting shows |
|---|---|
| Change | Replaced it with customer reviews. |
| Reported result | +11% product-page visits, +9% sales |
| Evidence | VWO case study, no baseline/significance disclosed |
| Limits | Both visits and sales rose after the switch. |
FAQ
What did Bunzl change on its homepage?
Bunzl replaced a crowded category section beneath the main banner with customer testimonials on Verzendverpakkingenshop, its online packaging store. Heatmaps and clickmaps had shown heavy interaction but scattered attention, and the team suspected the choices were slowing product exploration.
What results did VWO report?
VWO reports that the eight week test increased product page visits by 11 percent and total sales by 9 percent. VWO supplied the testing platform and published the case, so the lifts have not been independently verified.
Does the result prove reviews beat category navigation?
No. The test supports a narrow conclusion for this site, and the source discloses no baseline sales rate, traffic allocation, sample size, confidence interval or significance. Because the variation changed content and purpose together, it cannot separate fewer options from the reviews themselves.
How can another retailer reproduce Bunzl’s logic?
Find a high traffic module that gets many clicks but sends few visitors into product evaluation, keep one control, and replace the competing choices with one form of reassurance in the treatment. Measure completed orders or contribution margin per eligible visitor, and treat product page visits as a diagnostic.
Source note The experiment mechanics and figures come from a VWO customer case study checked on September 19, 2026. VWO supplied the testing platform and published the account. The source identifies Hans Breeuwsma as Bunzl’s conversion specialist but does not disclose the statistical details needed to independently assess the reported lifts.
Spotted an error? Contact the desk. Our corrections policy explains how we handle corrections and updates.
Find the reported method, result and limitations of other experiments in the Growth Signal Index in IC Lab.
