From analyst to running a $9M ad budget: ROAS up 230% by selling rooms, not categories
Ten of eleven marketers left within a short window. I had two weeks to learn a $3.5M paid program, then grew it to nearly $9M while making every dollar work harder.
BuildDirectDigital Marketing AnalystMar 2016 – Dec 20193 min read- +230%
- Return on ad spend
- $3.5M$9M
- Annual paid budget
- 2 weeks
- To take over the program
Context
BuildDirect was an online marketplace for building and home improvement products, sold to homeowners and to trade buyers such as contractors and architects. Flooring was the core, but the catalogue covered almost everything needed to build and furnish a home once the foundation was poured.
I joined as a digital marketing analyst. Then 10 of the 11 people on the marketing team turned over. Someone had to keep a $3.5M annual ad program running, so I stepped up, learned to run it within two weeks, and moved onto the marketing team to lead business analytics and all paid marketing. I reported to the Director of Marketing, and later to the CMO.
The thesisNobody shops for a product category. They shop for a kitchen, a basement, a floor that survives a dog. Sell the room and the job, and the product follows.
How the budget was split
| Share | Purpose | Where it ran |
|---|---|---|
| 60% | Prospecting: reaching new buyers | All major ad platforms |
| 10% | Message testing for in-funnel prospects, second purchases and win-back | All major ad platforms |
| 30% | Retargeting and remarketing | Major ad platforms and Criteo |
The budget was deliberately split by job. Most of it went to finding new customers. A smaller, fixed share was reserved for testing, so we were always learning which messages worked for people already in the funnel, for a second purchase, or to win back lapsed customers. The rest followed up with people who had already shown interest.
Ring-fencing the testing budget mattered. Without it, testing is the first thing cut when results dip, which is exactly when you most need to learn.
The testing loop
The three parts of the budget fed each other. Prospecting filled the funnel, the testing budget found the angles that moved people, and retargeting closed the loop. Whatever won in testing was rolled back into prospecting, so the whole program improved continuously instead of in occasional overhauls.
Selling rooms, not categories
Illustrative examples of the shift.
| Lead with | Category first | Room, use case or feature first |
|---|---|---|
| Room | Shop tile | Everything for your bathroom remodel |
| Use case | Shop vinyl flooring | Floors built for kids and pets |
| Feature | Shop laminate | Waterproof floors for basements |
The biggest lesson from the testing budget was about framing. Ads that led with a product category performed worse than ads that led with the room being built, the job to be done, or a feature that solved a specific problem. Buyers think in projects, not catalogue structures.
Results
Over the following years, annual paid spend grew from $3.5M to nearly $9M, and return on ad spend rose by 230%. Growing spend usually pushes returns down, because each extra dollar reaches less likely buyers. Better framing and continuous testing let us scale both at once.