FoundIt! makes Constructor better
Constructor optimizes product discovery against real commercial outcomes, learning continuously from clickstream and behavior. That learning is powerful, and is bounded by one thing: it can only learn from products customers were able to find and click in the first place. FoundIt! reads demand signals from outside and inside your business, so new products and lower-traffic categories arrive understood rather than waiting to accumulate clicks. More of your range becomes discoverable, clickable, and therefore optimizable.
The behavioral ceiling
Behavioral AI can only learn from what shoppers can reach
Clickstream-driven optimization is a genuinely strong approach. Show shoppers products, watch what converts, and let the system get better at showing the right things. Constructor does this well.
But there is a quiet ceiling built into any behavioral model. If a product is described as “sandals, brown, high heel,” it never surfaces for “espadrille wedges for a garden party.” Nobody sees it. Nobody clicks it. The model concludes, reasonably, that it isn’t worth surfacing, and the product stays invisible. The absence of a signal is read as the absence of demand.
Meanwhile, the demand was real. It just never reached a product that could satisfy it. Enriching the catalog breaks that loop and sits upstream of any optimization engine.
Side by side
Constructor alone vs. Constructor with FoundIt!
| Capability | Constructor alone | Constructor + FoundIt! |
|---|---|---|
Optimization pool | Optimizes across products shoppers can currently find | Widens the pool, so more of the range enters the optimization loop |
Long-tail products | Limited signal on products that rarely surface | Long tail made findable in customer language, generating signal it never had |
Query understanding | Strong intent inference from behavior and context | Behavioral inference plus catalog data written in the same customer language |
Cold-start & new products | Needs behavioral data to accumulate before optimization sharpens | External demand signals mean new products are understood on arrival, before any clicks exist |
Attribute depth | Works with the attributes you supply | Deep occasion, use-case, and mission attributes across every product |
Setup | Platform you already run | Layers on top; no rip-and-replace, feeds the engine you already chose |
Low-traffic categories | Thinner signal means slower learning | Outside-in intent data fills the gap where internal behavior is sparse |
Where FoundIt! fits
One learns from behavior. One makes sure there’s behavior to learn from.
Constructor is the optimization engine, tuning discovery against revenue, conversion, and the KPIs you actually care about. FoundIt! is the intent layer that determines how much of your range is eligible for that optimization in the first place.
The two compound. Richer product data means more products surface, which means more behavioral signal, which means the optimization gets sharper across a wider range. Constructor decides what to show. FoundIt! makes sure everything worth showing can be found.
Proof in the real world
The demand was always there
Making the long tail findable, and profitable
Ironmongery Direct carries more than 20,000 products, including a deep long tail that rarely surfaced. By making every product findable in the language customers actually use, their most recent test delivered an 8% year-on-year revenue uplift, with margin holding firm and customer numbers growing. Fenwick, applying intent to navigation across a 200,000-page catalog, added 9% to revenue in six weeks.
“We used to be a catalog retailer that happened to have a website. Now we’re a fully digital retailer.”
Give your optimization engine more to optimize
You invested in Constructor to convert. FoundIt! makes sure it has your whole range to convert with, not just the part customers can already find. We’ll show you what intent-led data unlocks on your own catalog.