New product images: when they grew sales and when they didn't

On 19 Oct the growth agent put new images on 11 martini glass sets. In the 2 weeks after, their sales went from $264 a day to $360. In the same days, 8 products from the same 2 sellers with no new images fell from $750 a day to $551.

The gain came from more sessions, not a higher conversion rate: sessions went from 96 a day to 196. On 14 other products, it went the other way: conversion carried the gain.

Superfuel runs a growth agent on Amazon brands' ads, prices, listings and images, and logs every change with why, what it expects and when it checks. Here's which products gained after new images, which didn't, and what separated them.

What the agent did, and how many times

117 image changes on 33 products, between 19 Oct and 5 Nov. Almost every product got a whole new set, main image and gallery, the same day. Each gain came from one of two places: more sessions, or a higher conversion rate.

The worked example: 11 martini glass sets

These are plastic coupe glasses that look like real glassware, with stems that screw on. They come in sets of 12 at about $20 to $25 and sets of 24 at about $25 to $32, in clear, gold rim, pink and flower shapes.

19 Oct, images. The 11 sets sold $264 a day from about 96 sessions, at a 9.8% conversion rate. The agent replaced the main image on all 11 and added 7 gallery images to each. On the sets of 24, the main image shows the whole pack with a tag giving the count and size, like "24 Pack / 7 oz", so the count reads at thumbnail size in search. On the sets of 12, the main image shows twelve coupes with four filled (champagne, rosé, espresso martini, chocolate mousse), to show they work for drinks and dessert. Sales on the 11 went from $264 a day to $360.

The gallery captions answer shoppers' questions. One on the 7 oz gold rim set of 24 reads: "Friends toast with them and nothing comes apart: the stem twists on and stays on." The reason: parts falling apart was the top complaint in rival reviews.

21 Oct, price. It raised the 7 oz gold rim set of 24 from $29.99 to $31.99, so each set would keep about $2 more, since its page says BPA-free and most rival sets don't. Its sales still went from $41 a day to $136.

When new images grew sales: more sessions on a page that already converted

The clearest case is the 7 oz gold rim set of 24, which tripled its sales after its price went up. Its sessions went from 14 a day to 47, and its conversion rate held at about 9% (9.5% before, 9.0% after). It already converted well. The new main image brought more people to it.

Across the 3 sets of 24 that gained, sales went from $75 a day to $269 (one also had a 15% price cut). Their sessions went from 40 a day to 125, and 73 of those 85 extra sessions were organic.

Here's what we tested to explain why these gained and others didn't. How well a set converted before separated them most; the next section shows it.

Pack size. The 6 sets of 24 went from $199 a day to $314. The 5 sets of 12 went from $65 to $46.

Price. On the 5 sets where average selling price held or rose, sales went from $183 a day to $221.

The 2 discounted days. On 4 Nov, 2 of the sets of 24 sold 46 units at about $6 to $11 each. That day sits outside the 2-week window, so it's not in any number here. Leaving both sets out, the other 9 went from $104 a day to $188.

Stock and ads. No set ran out in the window. On the 3 sets that gained, ad clicks went from 9 a day to 21, while sessions went up by 85.

When new images didn't grow sales

• Under 10 sessions a day. The 5 sets of 12 had 2 to 8 sessions a day. Sales went from $65 a day to $46.
• Conversion rate under 2%. The 2 pink 9 oz sets of 24 converted at 1.4% and 1.7%. One went from 4.9 sessions a day to 12.6, and from $2 a day to $5.
• Low stock. The clear 7 oz set of 24 was running low, so the agent slowed its sales to make the last units keep more. Sales went from $120 a day to $36.
• Price raised to stretch stock. 7 more sets got new images on 26 and 27 Oct, 4 of them with price raises on 21 Oct to stretch low stock. They went from $589 a day to $313.

When new images grew sales: a higher conversion rate

On the Bible covers, sip-lid cups, a bottle drying rack and a wipes holder, sales went from $172 a day to $370 over 7 to 14 days, in Bible cover season. Sessions rose a little, from 207 a day to 254. Conversion rate went from 5.4% to 9.1%. Ad clicks rose too, and ad traffic usually converts lower, so the conversion gain isn't the ads.

The new images put the biggest shopper question near the front. On the Bible covers, image 2 says which Bible sizes fit. On the cups: "I can drink right from the lid; I don't need straws."

What works, what doesn't, and why

• New images grew sales where the page already converted and needed more sessions. The 7 oz gold rim set of 24, at a 9.5% conversion rate, went from 14 sessions a day to 47 and from $41 a day to $136, because a main image that shows the pack count at thumbnail size wins clicks in search, and a page that converts turns those sessions into orders.
• Images that answer the top shopper question first lifted conversion rate where it sat around 5%. On the other 14 products it went from 5.4% to 9.1%, because the shopper's main doubt (will my Bible fit, can I drink from the lid) is settled before they leave the page.
• They did little under 10 sessions a day. The sets of 12 fell from $65 a day to $46, because at 4 sessions a day there is too little traffic for any image to change the result.
• They did little where conversion rate was under 2%. The pink 9 oz sets of 24 went from $2 a day to $5, because more sessions on a page that rarely converts still bring few orders.

When your images changed, was it sessions or conversion that moved?