The cheaper bundle upsell that earned PaintNuts more

Only 5% of customers were taking the main upsell offer.

PaintNuts sets the standard for automotive touch-up paint, matching the exact factory colour to a customer’s vehicle registration. Alongside the paint itself, the range covers everything a good repair needs: primers, lacquers, scratch and scuff removers and touch-up pens. Once a customer adds a product to their basket, the site offers a one-click bundle upgrade to add those extras in a single tap. It’s a valuable moment in the journey, and one we knew could work harder.

Previous testing had already shown that the bundle offer was an important part of the page, but only around 5% of customers were taking it. We wanted to understand whether the problem was the way the bundle was presented or the products included in the bundle itself.
A previous test had given us a useful clue. Although that test lost overall on revenue per user and was not implemented, heatmaps showed that clicks on the bundle had increased by 33%. That suggested the bundle could be improved, but we still needed to understand whether better copy was enough or whether the offer itself needed to change.

We compared improving the existing bundle with making the bundle more relevant.

Door4 runs PaintNuts’ conversion optimisation programme: an ongoing cycle of hypotheses, tests and rollouts across the store’s key journeys. So rather than rebuild the page on a hunch, we set the two most likely explanations against each other and let the data settle it.

We looked at the products included in the bundle. The existing bundle included primer and lacquer, which are useful for some types of repair but not every customer needs them.
Scratch and Scuff Remover has a much broader use case, so we wanted to test whether replacing the existing bundle with a cheaper but more relevant combination would encourage more customers to upgrade.

We also wanted to revisit the copy from the previous test, where bundle clicks had increased but overall revenue had fallen. So we tested both questions separately.

Door4’s CRO team designed and ran the test, agreeing the variants and success measures with PaintNuts’ stakeholders before anything went live.

We then ran a three-way test on the upsell page. The test included:

• Control: the existing bundle and existing copy.
• Variant 1: a new cheaper product bundle.
• Variant 2: the existing bundle with the stronger copy from the previous test.

Variant 1 changed only the products included in the bundle. Variant 2 changed only the copy, allowing us to understand whether the messaging from the previous test had been beneficial when tested independently.

The test ran for 14 days on the upsell page. We measured revenue per user as the primary metric, alongside conversion rate, average order value and the percentage of customers taking the bundle.

The winning version needed to generate more revenue per visitor.

Revenue per user was the primary measure because the new bundle was deliberately cheaper than the existing offer. This meant average order value could fall even if the test was commercially successful.

We therefore looked at:
Revenue per user — the main commercial measure.
Bundle upgrade rate — to see whether more customers found the offer relevant.
Conversion rate — to understand any wider impact on completed orders.
Average order value — to check the effect of the lower-priced bundle.

The main question was whether the increased number of customers taking the bundle would more than compensate for its lower price.

The new bundle nearly doubled the upgrade rate and increased revenue per user by 8.3%.

The new bundle produced a clear improvement:
Bundle upgrade rate: 5.0% → 9.6%
Conversion rate: +8.8%
Revenue per user: +8.3%

The fall in average order value was expected because the new bundle was cheaper. However, substantially more customers took the offer, which resulted in higher revenue overall.

The copy-only version also improved performance. Bundle uptake increased to 6.8%, while revenue per user increased compared with the control.
This showed that both the offer and the messaging had room for improvement, but changing the products included in the bundle had the larger impact. The winning bundle was subsequently implemented on the upsell page.

The test also gave us a useful principle for future upsell testing: the highest-value upsell is not necessarily the highest-priced one. A lower-priced product that is relevant to more customers can generate more revenue overall. The next logical test is to combine the winning bundle with the stronger copy and see whether the two improvements work together.

That principle is exactly what an ongoing testing programme is for. Door4 runs PaintNuts’ conversion optimisation programme, and each experiment like this one banks a repeatable insight the store keeps drawing on while pointing to the next test worth running. A lift of 8.3% in revenue per user is modest on its own, but compounded across the journey, gains like it add up to meaningful growth without a penny more on ad spend.