Koca Ventures Ltd
71-75 Shelton Street
Covent Garden, London
WC2H 9JQ, United Kingdom
Registered in England & Wales — 16231043
PRICE INTELLIGENCE
A price-intelligence pipeline you own —your sources, your catalogue, your systems.
Not a per-SKU SaaS seat you rent, and not a black box you can't leave. We build a private pipeline that scrapes the exact public sources you care about, matches them to your real catalogue with confidence scores, and wires the result straight into your own systems — on a public-data-only legal posture we lead with, not bury.
▶Product walkthrough — a live price monitor firing an alert, and a matched-product confidence view (coming soon)
HOW WE WORK — LEAD WITH THIS
Our legal & ethical stance is the feature
Most price-monitoring marketing glosses over how the data is collected. We put it first. We promise defensible, public-data monitoring — not “scrape anything.” Here is exactly what we will and won't do.
✓We collect public, logged-out prices and availability.
We don't scrape behind logins or paywalls.
✓We respect robots.txt and rate-limit politely.
We don't defeat anti-bot, CAPTCHA, or access-control systems.
✓No fake accounts, no credential use.
We don't impersonate users or platforms.
✓We minimise and avoid personal data; GDPR/UK DPA-aware design.
This isn't legal advice — you get counsel for your jurisdiction and use.
✓Auditable collection and repricing decision logs.
You own your pricing strategy and competition-law compliance (MAP/RPM, algorithmic-pricing antitrust).
✓A pipeline and dataset you own — on your infrastructure if you want.
Coverage of any specific site is best-effort; hostile anti-bot targets may be declined on principle.
The bright line in 2026 is public, logged-out data versus anything else. Staying logged-out, respecting robots.txt, and never circumventing access controls is the defensible posture — and it's the one we build on. This is not legal advice; for any sizeable program we recommend jurisdiction-specific counsel.
USE CASES
Different buyers want different things
01
Retailer repricing
Daily-to-near-real-time competitor price and availability tracking, fed into repricing logic with hard margin floors. The goal is never the lowest price — it's the highest price that still wins, with a guardrail that holds at your true landed cost instead of racing a competitor down.
02
Brand MAP & channel-price monitoring
For brands and distributors: spot retailers advertising below your agreed floor, and watch channel pricing across markets to protect brand equity. (We monitor and report — in the EU/UK we frame this as price monitoring, not enforcement, because resale-price maintenance is restricted by competition law. The commercial decision stays yours.)
03
Marketplace Buy-Box & assortment-gap
Stock and promotion visibility plus assortment-gap analysis — what competitors stock that you don't — and the signals that move Buy-Box wins. Structured so a category team can act on it, not just admire a chart.
04
Data-team structured feeds
If you have your own pricing or data team, you usually want the pipeline, not a dashboard. Clean schemas, stable structured feeds, and APIs that drop straight into your ERP, PIM, BI, or your own models — output you own, not a vendor export.
HOW IT WORKS
The five-layer pipeline
01
Collection
Headless browser drivers (Playwright, including Playwright-over-CDP) read public, logged-out pages. Vanilla drivers now get blocked on hard targets, so we apply driver-level stealth — the honest kind that makes a polite reader not look like a buggy bot, not the kind that defeats login walls or anti-bot challenges. If a site actively blocks automated reading, we read that as a 'do not collect' signal and respect it.
02
Network & proxies
Residential and ISP proxies geo-targeted to the markets whose prices actually matter, sized to the volume and difficulty of the targets. This is a real pass-through cost that scales with what you ask for — we're transparent about it rather than burying it in a flat seat price.
03
Anti-bot-aware reliability
For permissible public targets where a do-it-yourself setup hits a reliability ceiling, we use managed collection APIs — for uptime on pages we're allowed to read, never to beat an access-control system. The harder a wall a site puts up, the more it signals the operator doesn't want automated collection, and we treat that as a line to respect.
04
Product matching — confidence + human QA
The make-or-break layer. Where a GTIN/EAN/UPC barcode is present and correct, matching is a clean database join. The engineering value is the rest: fuzzy text matching, image and multimodal matching, and LLM semantic matching — each producing a confidence score per matched pair, with low-confidence pairs escalated to a human-QA loop. We promise confidence scores and a review loop, not an unverifiable headline accuracy figure.
05
Intelligence, repricing & dashboards
Change detection is not intelligence. We turn 'the number moved' into 'what it means' — vs your price, who leads the category, promo vs permanent vs pricing error — and feed repricing logic that is hybrid by design: transparent rules for the boundaries (cost-plus floors, inventory thresholds) and ML within those guardrails. Every repricing decision is logged for audit.
🖼Pipeline architecture diagram — collection to matching to repricing (coming soon)
Be honest about the cat-and-mouse: scrapers are a living system, not set-and-forget. Targets change layouts and defences constantly, so part of the work is perpetual maintenance — which is why the operate phase is a retainer, not a one-time deliverable.
WHY A CUSTOM BUILD
What an owned pipeline gives you that a SaaS seat can't
01
Bespoke sources
We scrape the exact regional, long-tail, and niche competitors and portals you care about — the ones the off-the-shelf vendors don't already cover. (We've already built a six-portal multi-source scraper for a different vertical; see below.)
02
You own the pipeline and the data
No per-SKU seat tax, no data held hostage. The codebase and the matched dataset are your asset, on your own infrastructure if you want it. Off-the-shelf platforms sell you a slice and keep the rest behind their wall.
03
Native integration
Output goes straight into your ERP, PIM, Shopify, repricing logic, or BI as clean structured feeds and APIs — not just another dashboard to log into.
04
Matching you can trust
Per-pair confidence scores and a human-QA loop instead of an unverifiable headline accuracy number. Accuracy is catalogue-conditional: a vendor's figure on their reference catalogue does not transfer to yours, so we don't pretend it does.
05
Vendor-risk insulation
A SaaS vendor can change its roadmap, get acquired, or file for bankruptcy out from under you — Wiser Solutions filed Chapter 11 in 2026 (public record). An owned pipeline can't go bankrupt on you.
PROOF — BUILT AND RUN BY US
A real multi-source scraper, retargeted at retail
We already run a production multi-source market-intelligence scraper for a client in another vertical — six public property portals across the EU, with geographic and market rollups and an anti-hallucination “honest-framing” guardrail that blocks overclaiming before it leaves the system. It's a demonstrated, real pipeline. The same engine, retargeted at retail prices, is exactly what this service is.
QUESTIONS
Straight answers
Is this legal?
We collect public, logged-out data only, respect robots.txt, rate-limit politely, create no fake accounts, defeat no login walls or anti-bot systems, and design to minimise and avoid personal data. That public-data posture is the defensible mode and the line we hold. This is not legal advice, though — every program is different, and for anything sizeable we recommend you get jurisdiction-specific counsel for your use. We build the engine and the audit trail on a careful posture; the legal sign-off for your program is yours.
Can you guarantee you'll scrape site X?
No honest shop can. Coverage of any specific site is best-effort — targets change their layouts and defences constantly, so monitoring is a living system, not set-and-forget. And if a site puts up a hostile anti-bot wall, that's a signal the operator doesn't want automated collection, which we respect on principle and will decline rather than try to beat. We're transparent about what we can and can't reliably read before you commit.
What accuracy do you promise?
We promise per-pair confidence scores and a human-QA loop on low-confidence matches — not a magic universal accuracy percentage. Headline accuracy figures are catalogue-conditional: a number measured on a vendor's reference catalogue doesn't carry over to yours. Where a correct GTIN/EAN/UPC barcode exists, matching is exact; the real engineering value is what happens the rest of the time, and we'd rather show you a calibrated confidence score than sell you a number we can't stand behind.
Do we own it?
Yes. The codebase, the schemas, and the matched dataset are your asset — and the pipeline can run on your own infrastructure if you want it to. There's no per-SKU seat tax and no data lock-in. The whole point of a custom build over a SaaS seat is that you keep it.
Do you set our prices, or handle MAP enforcement?
No. We build the engine, the matching, and the audit trail; your pricing strategy stays yours. On the brand side, in the US 'MAP enforcement' is a recognised concept, but in the EU/UK resale-price maintenance is restricted by competition law — so we frame brand-side work as channel-price monitoring and reporting, not enforcement. Algorithmic and AI-driven repricing also draws growing antitrust scrutiny; we keep auditable records of every repricing decision and treat the pricing rules themselves as your commercial and legal responsibility.
How do you price it, and can it run on our infrastructure?
Per engagement, after we scope it — there's no per-seat list price here. A typical path is a fixed-fee scoping pilot run on your own SKU catalogue (a real-catalogue pilot, because demos hide live-data complexity), then a project-fee build, then an optional monthly retainer to keep the scrapers healthy as targets change. Yes, it can run on your own infrastructure. Tell us the sources and the catalogue and we'll scope it honestly — proxy and anti-bot infrastructure is a real pass-through cost we'll be transparent about, since it scales with volume and target difficulty.
Last reviewed:
READY TO TALK?
Scope a pilot on your own catalogue
The honest first step isn't a demo — it's a real pilot on your own SKUs, because live data is where the complexity hides. Tell us the sources you care about and the catalogue you want matched, and we'll scope coverage, confidence, and a build before any larger commitment. Pricing is per engagement, after we understand the work.