Replacing your site search again? Why the problem usually isn't the engine but the data and intent behind it, and three things to do before you switch.
Let me ask you something uncomfortable.
This probably isn't the first time you've thought about replacing your search engine.
And if you're honest with yourself, it won't be the last. Unless something fundamentally changes.
I've watched a pattern play out across dozens of large retailers over the past decade.
Search starts letting customers down. Results are close but not quite right. Zero-result pages creep up. Conversion from search drops.
Someone raises the red flag. A tender goes out. Six to twelve months of integration work. Half a million to a million pounds. A new engine gets bolted in.
And for a while - it feels better.
Then, about two or three years later? You’re having the same conversation.
I've watched retailers cycle through legacy platform search, from SOLR, to Endeca, to Algolia, and on to whatever the new shiny thing is. Each time convinced that this is the one that fixes it.
It's not.
The Problem Isn't Your Search Engine
Here's what nobody wants to say out loud in those tender meetings.
The problem is rarely the search engine.
The problem is what you're giving it to work with.
Think of your search engine like a brilliant musician. It can only play the notes you've written down for it. If your product data is thin - organised for your catalogue, not your customers - your search engine is working with sheet music full of gaps.
It doesn't matter how sophisticated the technology is.
If a customer types "beach wedding guest dress" and you haven't enriched your products with that language, your search returns nothing useful.
Or worse, a page full of swimwear.
The search engine didn't fail. You gave it the wrong notes to play.
The Misdiagnosis That's Costing You Millions
We here at FoundIt! recently spoke with a retailer who was midway through a plan to rip and replace their existing search engine. Half a million pounds. Six to twelve months to deliver.
When we walked through what was actually happening - why search was failing - it became clear pretty quickly. Their intent coverage was thin. Products weren't enriched with the language customers actually use. The engine was doing its job. The data wasn't.
Replacing the engine wouldn't fix that. They'd spend all that money, go through all that pain - and within eighteen months, the same conversation would be happening again.
There's a medical analogy here I keep coming back to.
You wouldn't treat a caffeine headache with stronger painkillers.
You'd cut the coffee.
The same logic applies here.
Symptom. Root cause. Two very different things.
Retailers have been treating the symptom for years. What needs to change is the diagnosis.
The Two-Part Problem Nobody Joins Up
Here's something that's important to understand. There are actually two distinct failure modes happening simultaneously - and most retailers don't see them as connected.
Part one: Discovery failure.
Customers can't find you in the first place. They search externally - on Google, through social, through an agent - using their language. But your product content is organised in catalogue-speak. You're invisible to them.
Part two: Experience failure.
Customers do find their way to your site. They search. But search just drops them in the middle of a city they don't know, with 2,000 products and no map. Search didn't narrow the field. It overwhelmed them.
And they leave.
Two different failures. Same root cause - product content that doesn't reflect how customers think and talk about what they're trying to buy.
Fix the root cause, and you fix both.
The Future-Proofing Question Nobody's Asking
Nobody knows exactly how retail discovery will evolve over the next twelve to twenty-four months.
It could go fully agentic. It could be a blend of on-site search, off-site discovery, and conversational interfaces. It could be something none of us has quite imagined yet.
Here's what I'm certain of: locking yourself into a big new platform - with a six to twelve month integration timeline - is a significant risk when the landscape is moving this fast.
But here's what's also true. If you fix the core - if you enrich your product content with the language your customers actually use, structured as proper intent-led attributes - that intelligence travels with you.
Your existing platforms become better immediately. New platforms can latch on cleanly. Agentic layers can access it. SEO improves organically.
One of our clients, B&Q, took the intent data we'd generated back into their source data. It became their data. Their IP. Their asset to use however the future demands. They're not dependent on any vendor's black box. They own the intelligence.
That's the smart move. Not another platform bet. Building the core that powers everything - now, and whatever comes next.
Three Things Every E-Commerce Director Should Do Right Now
1. Audit your intent coverage before you look at any technology.
Before you commission another platform review or write another tender brief, map where your product data is failing your customers. Where are customers using language you haven't anticipated? Where does search return nothing, or something irrelevant? This is your real problem statement. And it might save you half a million pounds.
2. Ask a different diagnostic question.
Stop asking "Is our search engine good enough?" and start asking "Are we giving our search engine what it needs to succeed?" The two questions lead to very different answers - and very different investments. One takes six months and millions. The other can start showing results in weeks.
3. Think about portability, not dependency.
Whatever you do next, make sure the intelligence you build is yours. Not locked into a vendor's black box. Not dependent on a platform you might want to move away from in three years. Own the data. Own the intent layer. That's what makes you genuinely future-ready - not which platform you're on.
The retailers winning right now aren't the ones with the most sophisticated search technology. They're the ones who understand their customers well enough that any technology they use performs brilliantly.
Fix the core. Everything else follows.
Warren.
Warren Cowan is CEO of FoundIt! , the intent-led commerce platform for large retailers. FoundIt! helps brands like M&S, B&Q, John Lewis. Saks Global, Neiman Marcus, and Michael Kors make every product findable by closing the gap between how customers search and how catalogues are organised.