You're sick of looking at the website. I get it. You've seen the same homepage for years and you can already picture something better.
But being bored with it and it failing your customers aren't necessarily the same problem. Before replacing the lot, find out which one you're dealing with.
Perhaps the business has changed. Perhaps people can't work out which service they need. Perhaps updating a paragraph has become an afternoon's work. Any of those is a better starting point than “we need it to look modern”.
Write the complaint down properly
What is actually happening?
“People keep calling because they can't find the delivery information” is something you can investigate. “It needs more impact” needs another conversation.
Speak to the people answering enquiries, handling orders and updating the site. Ask what they keep having to explain or fix. Then ask customers what they came to do and where they got stuck. Let them describe it without steering them towards your preferred redesign.
You might uncover a design problem, missing information or a process that breaks after the form has already been submitted. They're all worth solving. They don't all require a new homepage.
Use the numbers to ask better questions
Analytics can show you where to look. It can't tell you what everyone was thinking.
A short visit could mean someone found the answer immediately. It could also mean they gave up. A form abandonment could be a technical fault or somebody deciding they weren't ready. Go and test the task before assigning the dashboard a personality.
Use the information you're permitted to collect, alongside conversations and observation. Watch somebody find the service and make an enquiry. Keep your mouth shut while they do it; explaining the page for them defeats the point.
For a low-traffic business, that can be a more useful starting exercise than an A/B test that never gets enough activity to tell you much.
Don't throw away the parts people use
Make an inventory before changing things. Important pages, downloads, links, content and integrations all need a decision.
Some of the old work may be doing its job very well. Keep it. A redesign doesn't need to prove its value by making every corner unrecognisable.
Preserve useful addresses where you can. When a URL changes, send its old address to the relevant replacement and check the redirect. Sending every old page to the homepage isn't much help to someone who followed a link for something specific.1
The same applies to familiar tasks. Change an awkward process when there's a reason, but don't make returning customers learn a new arrangement solely because the design team wanted a fresh start.
Choose the right amount of work
An unclear offer may need new copy and a better structure. A business that's changed direction may need the wider identity reconsidered. A site that's difficult to maintain may need rebuilding underneath a design people already understand.
I'd decide that before choosing a platform or getting attached to a visual concept.
A complete rebuild can be the right investment. It can also be a large answer to a fairly small problem. Ask what each part of the proposed work fixes and what you're choosing to preserve.
Agree what you'll check afterwards
Be specific about the improvement you're after. Customers finding the right service. Reliable mobile checkout. Enquiries reaching the correct person. A team that can publish without needing help for every change.
Record the starting position and agree how you'll check the result. Don't let the new screenshot become the entire evidence that the project worked.
Plan the changeover too. Test the content, forms and redirects; protect the existing data; know who approves the release and how to recover if it fails. Then check the actual live website. A deployment saying “success” doesn't tell you that a customer can complete the task.
A new site can look fantastic. Just make sure that, when somebody asks why you replaced the old one, you can point to something better than a different shade of blue.
