Case study · Visual brand intelligence · Ecommerce
High-resolution ecommerce image intelligence, delivered to Fortune 500 retailers.
A visual brand intelligence platform serving Fortune 500 retailers replaced internal image-scraping infrastructure with a managed pipeline — structured metadata, per-URL pricing, confirmed SLAs.
Engagement summary
Primary-image scraping + structured metadata across marketplaces and DTC.
The client came to us through direct referral. They did not want another data platform to administer; they wanted data delivered on a schedule, with a human on the other end when something changed.
“We are looking for services, not a tool to use ourselves.” — Senior Director GTM, visual brand intelligence firm
The problem
Fortune 500 retailers expect turnaround times internal scrapers can't hit.
Visual brand intelligence lives or dies by image quality and delivery speed. If a retailer runs a visual analysis on a catalogue that's six weeks stale, the insight has already expired.
The client's engineering team had built an in-house image pipeline, but every new marketplace meant another set of anti-bot quirks, another pagination pattern, another consent-banner workaround. The team that should have been shipping intelligence was firefighting scrapers.
The shortlist of outsourcing options was thin. Self-serve scraping APIs would have required their team to own the orchestration and the QA layer. Freelancer pipelines came with no SLA.
They needed a managed delivery model — data in, on a schedule, with someone to call when a target site added a CAPTCHA yesterday.
The solution
Managed image extraction pipeline with per-URL schema and confirmed turnaround.
We run primary-image scraping across the client's target marketplaces and DTC channels. Each URL gets pulled through a managed browser fleet, images are extracted in full resolution, de-duplicated (same image under different CDN paths), and normalized to a consistent format.
Structured metadata — product title, marketplace price at capture time, listing URL, position within listing — goes into a JSON manifest delivered alongside the image bundle.
Each campaign has a per-URL price, a confirmed delivery window, and a named point of contact. When a target site changes layout overnight, we patch first, tell the client second. Their team never sees the mess.
VISUAL INTELLIGENCE PIPELINE — PER-CAMPAIGN DELIVERY
┌───────────────┐ ┌───────────────┐ ┌────────────────┐
│ Campaign URL │──▶──│ Browser fleet │──▶──│ Image capture │
│ list (client) │ │ (proxies+CAP) │ │ + DOM metadata │
└───────────────┘ └───────────────┘ └────────┬───────┘
│
▼
┌───────────────┐ ┌───────────────┐ ┌────────────────┐
│ Bundle + JSON │──◀──│ Dedupe + │──◀──│ Full-res image │
│ to client DL │ │ normalize │ │ extraction │
└───────────────┘ └───────────────┘ └────────────────┘
Per-URL pricing · Confirmed SLA · Named point of contact The numbers
What this looks like in production.
-
Active since
2024
rolling recurring engagement, expanded multiple times.
-
Delivery model
Per-URL
pricing with confirmed turnaround window per campaign batch.
-
Ops footprint
Zero
scraping infrastructure the client maintains in-house.
// this pattern repeats
If your team spends more time maintaining scrapers than shipping intelligence — there is a better division of labour.
Per-URL pricing. Named point of contact. No tools for your team to administer. Quote in 24 hours.