A kitchen manager places an order on Monday for Wednesday's truck. One item is sitting in a different distribution center, so it won't be on that truck. Nothing in the ordering platform tells her. She finds out on Wednesday, when the truck shows up without it.
My team’s research found that problem and wrote it up. Then we decided to leave it alone.
During those same quarters, my team was moving customer accounts onto that platform at 20% a quarter. Leaving the truck problem alone was part of how we did it.
The target
I was the UX manager for Gordon Food Service’s $4 billion B2B ordering platform. My team did all the UX work on it: research, workflows, the design system, component design, analytics, user testing, and documentation.
Our target was to grow the number of customer accounts ordering online by 20% every quarter. Each quarter’s goal started from the last quarter’s total, so every quarter we needed more new accounts than the one before. Over three quarters that compounds to about 73%. We hit it all three times.
Nearest barrier first
With that target, setting priorities was easy. Whatever stood between a customer account and its first online order came first: account selection, self-service password recovery, online order history, payment options.
Cross-docking waited. Getting a product from one distribution center to another in time for Wednesday’s truck, or at least telling the customer it would come later, was a real problem for people already ordering online. Solving it wouldn’t have brought a new account on board, so we held it. It was also hard. Gordon Food Service was a $16 billion company then, and cross-docking touched data, physical infrastructure, teams, and system integrations all along the distribution chain.
That clarity was good for my designers. They knew the work in front of them would bring customers onto the platform, and that gave them confidence.
What kept the investment coming
UX sat inside the Digital Experience team. Its leader was often brought into executive meetings to report on progress. She could point to outcomes from work we had already shipped and to specific work in the pipeline aimed at adoption, and leadership kept investing.
She also brought some of our research into those sessions. Leadership could see what was coming next and how it would bring more customers online, well before anyone asked them to fund it.
That pipeline carried us late in my year there, when months of work consolidating two technical environments finally landed. Very little of that work changed what a customer saw. Because the features we had queued up earlier were still releasing, we kept some momentum through it.
The cost of hitting it
Every account we brought online was one more customer using a system we knew would frustrate them.
Our own research showed it, and we had chosen the order of the work ourselves. Each quarter we hit the target, more people were living with the problems we had held, cross-docking among them. The designers felt it. So did I. As more customers came on board, all of us grew more anxious to start removing friction for the ones already there.
We also knew adoption could only climb so high before it stopped being the right thing to measure. When it did, we moved to measures of how well the platform served the customers already on it: order completion rates, basket size, the number of help and support tickets, and time to checkout. They’re standard ecommerce measures, and every one of them picks up the friction an adoption count can’t see.
What I’d recommend
If you lead design and someone hands you a growth target, take it, and use it to set the order of the work. We went after account selection and password recovery before anything else, and that is how we hit it.
Give your leaders the next quarter’s work along with this quarter’s results. Our Digital Experience leader did that every time she reported, and it’s a big part of why leadership kept investing.
Keep the work you hold in the backlog, and keep arguing for it. We held cross-docking on purpose, and we kept pushing for it to make a sprint, but in a company that size, a feature touching that many parts of the distribution chain was a big lift.
Know what you’ll measure once the target tops out, and have it ready before you need it. Ours put the frustration we had been watching grow onto the same reports leadership already read.
The kitchen manager counted toward our 20%. Her item still wasn’t on Wednesday’s truck.
