Start with an audit, not a redesign
A redesign is the most expensive way to find out your product was mostly fine. An audit costs a fraction, takes days, and usually tells you the problem is three specific things rather than everything.
Something is wrong with your product. People are dropping out, support keeps answering the same question, or it just feels dated next to a competitor. The instinct is to redesign it.
That instinct is usually wrong, and it is expensive to be wrong about.
What a redesign actually risks
A redesign replaces everything, including the parts that were working. You will spend months and a lot of money, and at the end you will not know which change helped, because you changed all of them at once.
It also resets what people know. Every user who had learned where things are has to learn again, and some of them will not bother.
None of that means redesigns are never right. It means they are a big bet, and a big bet should come after a diagnosis rather than instead of one.
What an audit is
Somebody goes through the product the way your users do, writes down what is wrong, and ranks it.
That is it. No new screens. The output is a list, and the list is the value.
I did exactly this on a browser-based BIM viewer, and the finding that mattered was not a missing feature. Five different colour systems were in play, three of them using the same three hues, so red meant a late task, a deleted element, or the X axis depending on a mode you could not see. Elements with no data were drawn white on a near-white canvas, which made the things you were meant to ignore the brightest on screen. That last one was a single constant in a handful of files, and it was the largest visible improvement in the product. The whole audit is written up.
Nobody would have found that by redesigning. A redesign would have picked new colours and rebuilt the same collision.
The ranking is the hard part
Anyone can produce a list of fifty things wrong with a product. That list is useless, because everything on it looks urgent and nothing gets done.
The method I use is to rank by how much pain a thing causes, multiplied by how often anyone hits it, divided by the effort to fix it. Dividing by effort is what makes the list actionable. It surfaces the small, cheap, high-impact things that a severity-only list buries under the big ones.
On that audit, the single highest-ranked item was a colour constant. The most dramatic-sounding finding, a whole feature that was a placeholder, ranked lower because it was a month of work.
What you should get
- A ranked list, with the reason each thing is where it is
- Evidence for each finding, not an opinion. A measurement, a screenshot, a flow that fails
- What to fix first, and what to leave alone
- An honest statement of what was not checked, because a day of auditing cannot cover everything
If you get a deck full of adjectives about how the product feels, you bought the wrong thing.
Measure, do not squint
Most audits are somebody looking at screens and having opinions. Opinions are cheap and arguable.
Things I have found by measuring that looking would have missed: a divider at 1.15:1 contrast that reads as "a bit faint"; a menu 133 pixels off its own axis that looks roughly aligned; a 20 pixel icon inside a set of 16 pixel ones that looks fine until it sits next to them; a grid cell wrapping into a hidden fourth column, which produced two unrelated-sounding complaints from one cause.
Every one of those is checkable. That is the difference between a finding and a preference.
When you should redesign anyway
Audits are not always the answer.
If the underlying model is wrong, no amount of fixing the surface helps. If your product has grown by accretion until the navigation reflects your org chart rather than anyone's task, you have a structural problem and patching it is throwing money away.
The audit will tell you that too. It is just a much cheaper way to find out than a redesign that stalls halfway.
The practical version
Ask for a few days of somebody's attention before you commit to months of it. You will get a ranked list, an idea of how they think, and a sample of what working with them is like, for a fraction of the cost of being wrong.
If that sounds useful, tell me what is going wrong and I will tell you whether an audit is the right first step or whether you already know enough to skip it.