Mekael Wesley-Rosa Product designer, Chicago

Redesigning the whole app without turning it off

First designer Runwise ever hired. I rebuilt nine screen areas in six months while thousands of buildings ran their heat through them, in the middle of heating season.

The overview I chose to rebuild first still carries 37% of every pageview in the product. I picked it on instinct and had no way to know.

RoleSole designer
TeamPM, CEO, 4 engineers, me
CompanyRunwise
ScopeEvery customer screen
Why the building overview had to be rebuilt first Nine screen areas were rebuilt over six months. The building overview went first because every screen below it inherits its structure, so getting it wrong would have meant redoing the other eight. The water tab, a screen that is empty for most customers, later sourced a closed deal. Building overview rebuilt first, on purpose Heating Water List Graphs Targets Alerts Push Settings Store Every one of these inherits the overview’s structure. Get it wrong and you draw all eight twice. The water tab is empty for most customers. It sourced a closed deal three weeks after launch. The overview I rebuilt first still carries 37% of every pageview in the product.
Sequencing was the only decision that could not be undone. The shaded screen is the one I would point at, and it is the one with nothing on it.

What I walked into

The old product worked. Real buildings stayed warm because of it.

Every screen was a reasonable answer to whatever question was in front of somebody that day. There were just no shared answers, because nobody’s job was to hold them.

You could read the order the features arrived in, right off the screens.

And this isn’t a website. A superintendent runs the heat in a building where people live, and it’s January.

So the constraint was never aesthetic. Somebody had to run their building on Tuesday exactly like Monday, without being told anything changed.

You don’t relaunch that. You replace it one piece at a time while it’s running.

My role

Only designer, working with a PM, four engineers, and our CEO, who reviewed most of it personally.

Not mine: the design system this work needed. The token work happened about a year later, done properly, by the team.

Overview first, on purpose

Nine screens, one designer, and only one decision I couldn’t take back.

Whatever structure the building overview landed on, the eight screens under it would inherit. Get it wrong and you draw all eight twice.

I couldn’t work out what that page wanted to be.

It was ranking the buildings you should look at first, and also telling you about the same things again in a list on the side.

Settling that one page took longer than drawing the next four combined. It feels like the only decision on this project that really mattered.

Nine rounds on a page with nothing on it

Most customers bought heating and only heating, so the water tab had nothing in it. Designers call that an empty state.

It’s the screen everybody hands off. Gray icon, “no data available,” go work on something real. I’d rather spend the time there.

I did nine rounds with engineering on it instead. What we landed on explained what water monitoring does for a building like theirs, and gave them one way to ask.

Three weeks after launch our head of product traced a closed deal back to a customer clicking that page.

I don’t think that was luck.

It’s the only screen where you already know what somebody doesn’t have, shown to the exact person who doesn’t have it, at the moment they went looking.

The convention is to put an apology there.

Introducing the company as my critique

No peer to review with and no critique to hold. So I posted everything into the product channel and let whoever was there be the critique.

  • Lee, our CEO, argued with me about how charts behave when you touch them.
  • Two engineers talked me out of my chart-hiding logic, and were right.
  • Field staff told me what installers would hit before one hit it.

None of that reaches you in a design review. I still work this way at a company that has design reviews now.

We rolled it out internally through a weekly Feature Friday email, one whole edition on the redesign.

A redesign nobody on support can explain is a support problem with better typography.

What it cost me

All nine shipped over six months with no interruption to any building. That was the hard requirement and it’s the thing I’m most confident about.

I also had no way to tell whether I made any of them better. Nobody had set the product up to count what people clicked, so there is no before.

I didn’t add counting as I went, because shipping felt more urgent. It always does.

That cost years. And it cuts both ways: I was right about the overview and I could as easily have been wrong and never found out.

So I write the counting into the plan now, before anybody reviews the design.

I’ve since argued a full year of my own work should be cancelled on numbers I went and pulled myself.

Nine screens, no baseline

Screen areas rebuilt9
Downtime during heating seasonNone
Deals sourced by an empty state1
Share of all pageviews on the overview, today37%
Before-and-after data on any of itNone