At eCommerce Expo in September I told a room full of eCommerce managers something blunt. If their agency needs a fortnight to say how the retainer hours get split, they’ve got a pretty big fucking problem. I stand by it. The talk’s below, all eighteen minutes of it, clicker tantrum included. Thanks to Hyvä and Mollie for putting me on the Connected Commerce Stage in the first place.
Two lists, and why the fires win
Every eCommerce manager I’ve met has two lists. The one on the left is what they actually want to do, so a better checkout, proper A/B tests, a product page that sells. The one on the right is the product feed that’s died again, the discount codes doing something weird and wonderful, and the patch that’s just knocked out half your extensions. I asked the room for hands on both and the right-hand list won easily.
It wins every time because fires are loud, they’re certain and they end. Karen’s message in capitals gets answered today, the ticket tells you exactly what to fix, and once it goes quiet it’s off your plate. Meanwhile the checkout you wanted to sort last quarter is still sat there costing you money.
Three questions separate the fires from the cause. Has it come back? Does it need somebody senior? Would it still trip up a new starter? Two yeses and you’re patching a symptom, and you’ll be patching it again next month.
What the fires were hiding at Helmet City and Willow & Hall
Helmet City came to us in 2023 because GA4 didn’t work. It didn’t work because the site was on a bloated theme dragging 40 or 50 plugins nobody needed (just for the shits and giggles, as far as I can tell). We fixed the tracking, and the first thing the data told us was that GA4 had never been the problem. The site was. We rebuilt it on Hyvä in seven weeks, right up against Black Friday, and took four seconds off the load time. Conversion went up 16% on the same platform, the same catalogue and the same back office.
Willow & Hall thought they had an abandoned cart problem. Their sofa and bed builder was the best thing on the site, but the journey around it was so full of friction that people would configure their perfect sofa and then leave at the point of purchase. All of it sat on Magento 1 years after it died, with Sage Pay support about to be switched off. We moved them to Magento 2 on Hyvä, rebuilt the whole experience and binned over 100,000 SKUs on the way. The site got 80% faster and sales went up 15%.
What happened to the retainer
Before Hyvä, both of them were burning at least 70% of their retainer on maintenance and fires. Now it’s the other way round, for the same money. Helmet City had wanted proper on-site search for years, and once the site stopped eating the hours we built it in a couple of days. Along with clearer sizing and smarter cross-sells, that’s added another 9% to conversion, and turnover’s up 5%. Willow & Hall change their configurator in days now, where it used to take months.
So ask your agency two things. How much of your time goes on maintaining the site and how much on improving it? If it takes them a fortnight to answer, you’re renting stability and you don’t own anything. And how long would it take to switch payment provider? We moved both of these to Mollie in under 12 hours with a 15-minute deploy (smoke duly blown up Mollie’s backside). If the answer’s months, you’ve got debt. If it’s never, you’re being held to ransom.
How to work out what better conversion is worth
Divide your target conversion rate by your current one, take away one, and multiply what’s left by your annual revenue. A £6m brand on 1.8% desktop and 1.1% mobile, with two thirds of its traffic on mobile, blends to about 1.33%. Drag mobile halfway to desktop and you’re at 1.57%, which is about a million quid a year more without another penny on ads. It stays on the table for as long as your team’s stuck on the right-hand list.
What comes after the fires
Get the fires out and you finally get to work the list on the left, and every result that comes off it hands you your next nice problem. Those are decisions you only get to make because something worked. You’re taking on more staff, or finally buying that expensive domain you’ve had your eye on. You’re expanding a product line, or bringing in a 3PL because you can’t keep up with demand any more. Nobody gets one while the site’s still setting itself on fire.
Getting clients there is the whole job, and it starts with the foundation, because speed and user experience are phase one of CRO. Bolt tests and tools onto a broken base and all they do is expose what the platform’s already doing. (Neither of these brands left Magento either. They fixed what they were standing on.)
Thanks again to Hyvä and Mollie, who’ve been great partners to work with these last few years, and to Helmet City and Willow & Hall for letting me pull them apart in public.
Get the slides and a fire log
If you want the slides or the fire log template from the talk, connect with me on LinkedIn and I’ll send them over. Fill the log in for a month, send it back, and I’ll tell you which three fires I’d put out for good first. If you’d rather start with your own numbers, our Free CRO Review runs your live speed data and shows you the first five issues.
