Turn Anonymous Amazon Buyers into a Direct Customer Base

The challenge

ColorCase builds custom Pelican cases, hand-assembled to order in colour combinations nobody else offers. The product sells itself. The problem was everything downstream of the sale.

Volume came through Amazon, and an Amazon sale is a stranger. The case ships, the customer disappears into Amazon’s ecosystem, and there is no name and no email. When a new colour launches, when Landing Gear wheels arrive in a new finish, when a TrekPak insert would finish somebody’s setup, there is no way to tell the person who already bought. Hundreds of customers a week walked away from a relationship the business had earned and could not keep.

Meanwhile the operation ran across five systems that did not talk: Amazon Seller Central, Shopify, the retail store, ShipStation for fulfilment and Fishbowl for inventory. Getting a picture of a weekend meant opening all of them. Running real lifecycle marketing on top of that, with abandoned carts, post-purchase follow-up, review requests and cross-sell campaigns, is a full marketing department’s job. The company did not have a marketing department.

What we built

A system that reads all five channels as one operation and acts on what it finds.

Recovering the customer. Amazon orders get matched against contact data, so a transaction that arrived anonymous becomes a named customer with a purchase history. That is the foundation everything else stands on, and it compounds: the contact database grows off the Amazon volume the business was already generating, without anyone doing data entry.

Selling to the customers already there. With purchase history attached, the obvious gaps become visible and addressable. Military-segment buyers who bought a Protector case but no wheels. Buyers who took a case with no custom foam, every one of them a TrekPak or divider upsell. The system segments the cohort, drafts the sequence in the brand’s voice, attaches a repeat-customer code, and flags prior purchasers for separate handling.

Running the DTC engine. Abandoned cart recovery, post-purchase review requests at day seven, accessory campaigns to owners of a given case who never bought the matching part, and social drafts queued for approval. Campaigns arrive as proposals with the segment already built and the projected reach attached, not as a task somebody has to remember.

Opening enterprise accounts. Military procurement, film production rental houses and industrial equipment companies buy standard Pelican in volume. The system identifies them from public contracting data and social signals, and drafts outreach shaped per segment: GSA compatibility and mil-spec colour options for military, a branded demo case offer for film, colour-coded asset identification for industrial. The value is the quarterly reorder cycle that follows a first order.

Holding the carriers honest. Negotiated shipping rates erode quietly. Carriers push a general rate increase every January, accessorial charges shift, and dimensional weight thresholds move. The system audits the bill against the agreement and assembles the analysis before a rate conversation, instead of after.

The operator drives all of it by voice memo from the assembly floor. Approve the launch post, send outreach to these two, hold the others, set up the carrier call for Wednesday. Six things move in thirty seconds and nobody opens a laptop.

What changed

The customer base compounds instead of leaking. Amazon volume now feeds the direct business rather than terminating in it, and the contact database grows without anyone doing data entry.

Marketing runs without a marketing hire. Lifecycle campaigns, segmentation and drafting happen continuously. The operator approves and redirects rather than building.

Enterprise became a motion rather than an accident. Prospects surface weekly with outreach drafted per segment, so the account type with the best reorder economics gets worked deliberately.

The weekend stops being a blind spot. Monday starts with what happened across every channel already reconciled, not with five logins.

Why NimbleBrain

The systems were not the problem and we did not replace any of them. Amazon, Shopify, ShipStation and Fishbowl each did their job; nothing read across them, and a person was the integration layer.

We connected what was there and put the operator in front of it through a voice memo, because the constraint on this business was never software. It was that every decision needed somebody to open four tabs first.