Dark Horse

noun: dark horse; plural noun: dark horses

A team, or entity that emerges to prominence or success unexpectedly or is not widely regarded as a favourite.

The Challenge:

Figuring out what to sell, to whom, at what time when working to a business-centric profit target.

What we solve:

We help businesses transcribe their inventory documentation into media plans with a statistical confidence in their revenue projections.

The Output:

A.I. inferred product categorisation in order to achieve profitable growth, revealing in some cases those products which might reveal themselves as dark horses.

The Process

  • Three professionals, two men and one woman, examining a line penalty document on a wooden desk, with computers and a lamp in a warmly lit office. The man on the left is pointing to the document, and all are focused on it.

    We start with a retailer’s line detail document — rows and rows of product data waiting to reveal opportunity.

  • Three business professionals examining a document in an office, with a glowing AI machine labeled "VERTEX AI" on the table in front of them. The AI displays search options including "finance report," "workflow," and "data trends."

    Using semantic recognition via Vertex AI, we align each product with real-world search demand. Think of it as fluent translation — from SKU-speak to search-speak.

  • Three scientists in lab coats conducting an experiment with beakers filled with glowing green liquid, labeled 'Margin headroom' and pouring into smaller containers. Two vintage computer terminals display text advertisements, with one saying 'Customer Acquisition Cost' and the other 'LTV MODEL.' A small black horse figurine is on the table in front of the computers. The scene resembles a retro or vintage laboratory setting.

    We calculate how much we can afford to spend to acquire a customer — based on each product’s intake margin.

  • People in lab coats and business attire are gathered around a conference table with scientific glassware emitting colored smoke. A holographic display shows a graph labeled 'ALLOWABLE COST X FORWARD VIEW,' with an upward trend line. A horse wearing a cape stands nearby.

    We assume future demand mirrors the past. That way, we model media spend across time — matching auction-based cost with seasonal appetite.

  • Three people playing a game on a spinning wheel labeled with features like 'Linear', 'U-Shaped', 'Time Decay', and 'Last-Click Attribution'. The wheel has multicolored sections and darts. A shadow of a horse is cast on the wall behind them.

    We map your responses to our questionnaire to the attribution model you feel most appropriate for your business.

  • Vintage scene of three people standing on a sidewalk near a sign that reads "Super Sale! New TV Sets". One man is pointing at the sign, another woman is looking at a flyer, and a third man is checking his portable radio. In the background, a life-sized model of a black horse with red mane and tail, and a white picket fence with a moving horse statue.

    Based on all these inputs and and the touch points you state you use, the final flourish is to assemble a media mix with a confidence level on the revenue that plan will generate.