Intent, structure, product data and channels: why fixing three of the four forces of product discovery still lets you down, and the questions to ask.
Why three out of four will always let you down.
I was on a call last week that kind of stopped me in my tracks.
Not because anything dramatic happened. Just because of a question someone asked that I've been thinking about ever since.
It was an e-commerce director at a large UK retailer, someone who's clearly switched on, clearly invested, clearly doing all the right things.
They said: "We've tried better search. We've improved our navigation. We've cleaned up some product data. Why is discovery still not working?"
And I thought, there it is. That question right there.
Because the answer isn't that any of those things were wrong. It's that they were disconnected.
They were treating product discovery like a list of separate problems, each with its own solution. Tick the box. Move on. Wonder why it still feels broken.
Sound familiar? I thought so.
The Four Forces Nobody Talks About Together
I've been wrestling with this thought for a while now, and I think I've finally got it to a place I can articulate properly.
Product discovery isn't actually a single thing. It's the output of four forces working together.
And the kicker is that you need all four. Not three. Not three and a bit. All four.
We call these the Four Forces of Product Discovery. Each one amplifies the next. Miss one, and the whole thing stalls.
Let me take you through them.
Force 1: Intent - The Understanding Layer
This is where everything starts.
Your customer doesn't search the way your catalogue is organised. They never have. They search the way they think. The way they feel. The way they talk to their friends.
"Something smart but not stuffy for a work do." "A drill that works for a first-time DIYer." "Curtains that go with a grey sofa."
That's intent. Contextual, human, often messy - but real.
The fundamental mistake most retailers make is building their product data around their internal logic. The buyer's taxonomy. The supplier's SKU structure. The warehouse system.
None of which has anything to do with how a customer thinks.
Getting Force 1 right means capturing those signals properly - not just a keyword here and a click there, sessions stitched together imperfectly across five tools that don't talk to each other. It means structured intent data: the relationships between what customers type, what they mean, what they want next, and what they eventually buy.
Every interaction sharpens the signal.
Get Force 1 right, and the rest becomes possible. Get it wrong, and no amount of downstream tooling compensates.
Force 2: Structure - The Customer Decision Tree Layer
And this is the one most retailers completely skip. Which is a big part of why things don't work.
Understanding what customers want is one thing. Turning that understanding into something your business can act on at scale - that's another thing entirely.
Structure is the bridge. It's what turns intent from a good idea into an operational framework.
Think about how customers actually make decisions in your categories. It's not linear. It's not flat. It's relational. Black connects to sequin connects to smart occasion. Quilted jacket, Barbour & waterproof all connect to a “Dog Walking”. These aren't tags, they're relationships. Hierarchies of what customers care about, and the semantic connections between them.
Without structure, intent never becomes operational. You understand what people want, but you have no consistent way to connect that understanding to your products, your navigation, your search, your feeds.
With structure, every product and every channel inherits the same underlying customer logic. The decision tree is built once, and everything downstream benefits from it.
This is the layer most platforms don't even have a name for. And it's the reason why so many intent initiatives deliver a spike and then plateau.
Force 3: Product Data - The Findability Layer
So now you understand intent, and you've structured that understanding into a customer decision framework.
Now you apply it. At scale. Across everything you sell.
This is where most retailers hit the wall. Their product data is still organised around catalogue logic, not customer logic. Thin attributes. Generic titles. Descriptions that tell you what a product is made of, but not what it's for.
The result? Your search engine can't find what the customer is looking for. Your filters don't reflect how anyone actually shops. Your Google Shopping feed is technically complete but contextually empty.
Intent-led product data means every product is enriched with structured attributes that reflect real customer missions. Tags that match how people search. Attributes that power filters customers actually use. Descriptions that connect product to purpose.
And critically, this doesn't mean a one-time enrichment project that decays the moment a new season arrives.
It means a living data layer that keeps itself current. New product in your range? Intent-attributed the moment it lands. Season shifts? The signals shift with it. No dark aisles. No lost SKUs. No inventory that only surfaces when a customer searches the exact keyword your merchandiser happened to think of.
Force 4: Channels and Platforms - The Distribution Layer
This is where it gets really interesting.
Because product discovery isn't just your website anymore.
It's your Site Search. Your navigation. Your PIM. Your PDPs. Your SEO. Your Google Shopping. Your PLP filters. Your AI answer engines. Your marketplace feeds. TikTok Shop. ChatGPT Commerce. All of it.
Every channel and platform through which a customer might find your products.
The question is: does your intent-enriched, structured product data actually reach all of those places - in the right format, at the right time, consistently?
Most retailers have data. They even have some intent understanding. But it stays locked in the core. It doesn't flow into every channel that determines whether a product gets found.
The part e-commerce tech misses entirely is that every channel is also a sensor. Every interaction on every surface feeds signals back into Force 1.
That's what makes it a flywheel, not a pipeline.
Three out of four feels like progress. It isn't. Not really.
Intent without structure? Great signals, no operational framework.
Structure without product data? A map with no territory.
Product data without channels? Intelligence that never leaves the building.
All four, pulling together? That's when the compounding starts.
The Flywheel Effect
So that's the four forces. But here's the part that I think is genuinely underappreciated — and this is what I want to spend a moment on.
When all four are connected and working properly, something starts to happen that's hard to explain but easy to feel once you've seen it.
Intent shapes structure. Better signals build a better customer decision tree. Structure organises product data - attributes, hierarchies, relationships applied at scale. Product data unlocks every channel, making every product findable on every surface. And every channel feeds more intent signals back in. Every click, anywhere, sharpens the understanding.
Round and round it goes.
Each cycle, your products become more findable. Each cycle, your data gets smarter. Each cycle, your channels perform better.
We call this the Intent Flywheel, a continuous pipeline, if you will. And the reason I love that framing is that it makes clear this isn't a project with a start date and an end date.
It's a compounding operating model.
When all four forces are pulling together, our retailers typically see a four to six per cent uplift in site-wide revenue. But honestly, the number isn't really the point. The point is that it keeps getting better.
Fenwick added nine per cent to revenue in six weeks. Not from a replatform - from applying intent signals to navigation. Neiman Marcus saw a 2.1 per cent uplift in conversion and a 2.2 per cent drop in bounce rate. 200,000 product pages optimised in 10 weeks.
Those aren't tool stories. They're flywheel stories.
The Bit That Scares People
I'm going to be straight with you here, because I think it's more useful than being vague.
Getting all the forces to work together requires going left.
Shift left. Back to the source. Back to how your products are understood, described, and structured at the core, before any of it reaches a channel.
That's not a quick win. It involves more people. It asks more of the business.
But it's the only fix that doesn't need to be redone in eighteen months. Every other approach is just polishing the output of data that was wrong to begin with.
And critically: you don't have to rip anything out.
That's actually one of the things I'm most proud of in how FoundIt! works. We layer onto whatever stack you've already got. Your search engine, your PIM - they stay. We sit underneath and around them, making sure all four forces are actually connected and pulling together.
Force 1 to Force 4, in sequence, compounding. That's the work. And it's where the flywheel starts to spin. Continuously.
Four Questions Worth Asking Your Team
You don't have to tackle all four forces at once. You just need to know which one is your weakest link, because that's the one holding you back.
1. Intent: Do we actually understand what customers want?
Not what you assume, what the signals show. Pull all search terms from all categories. Look at exit points. Map where customers are abandoning. Is that intelligence sitting in five different tools that don't talk to each other, or is it structured and usable?
If the answer is 'sort of', Force 1 isn't pulling its weight.
2. Structure: Have we turned that understanding into a decision framework?
Not flat tags, actual relationships. The hierarchy customers use to decide. The semantic connections between attributes in your key categories. The map of how people actually make decisions when they shop with you.
If that map doesn't exist, intent never becomes operational.
3. Product data: Can every product be found by every relevant intent?
Not just by the keywords your merchandiser anticipated. Take one category. Search and navigate to it the way a customer would - ten different ways. How many times does the right product surface? If it's not consistent, your findability layer has gaps.
Gaps in product data mean dark aisles. Dark aisles mean lost revenue.
4. Channels: Does your intelligence show up everywhere your customers do?
On-site search, navigation, SEO, answer engines, Google Shopping, marketplace feeds. Is the same intent logic powering all of them, or are they running on different logic built by different teams at different times?
If the answer is 'inconsistently' - Force 4 is where your flywheel is stalling.
Want to find your weakest force?
We work with multi-category retailers like Marks and Spencer, John Lewis and Kingfisher to map all four forces, identify exactly where the flywheel is stalling - and fix it without touching the rest of the stack.
DM me on LinkedIn, or visit foundit.com to start the conversation.