×Upspring
My creative strategy is only as good as what I know about the category and the creative we're running. Upspring's in every report I do: what to shift, what to keep, where the whitespace is. That's the difference between scaling creative and hitting a wall.Casey FallonCreative Director · brand.co
About brand.co
brand.co runs creative strategy for a portfolio of brands like Oats Overnight, Casper and Bezel. The job is to decide what to make next: which angles to push, which to retire, and where a category has left room nobody is using.
Those decisions rest on two bodies of knowledge. What their own creative is doing across every client, and what the rest of the category is running right now. Both have to be current, and both have to live somewhere other than one person's head.
Strategy is capped by what you can see.
A creative director can only reason over what they have actually looked at. Ads are video, images, copy and the pages behind the click, so looking at them means watching them, one at a time, per brand, every month.
It can't be searched or sorted the way a spreadsheet can.
Every brand on the roster is its own pass, from scratch, every month.
Monthly reporting consumes the same hours the thinking needs.
At one or two brands a strategist can absorb the cost. Across a portfolio it compounds, and what runs out first is attention.
Two connections, two bodies of knowledge.
The API carries Upspring's ad corpus, which is where the category work happens: what competitors are running right now, which patterns keep winning, how one advertiser's whole approach breaks down. The MCP connection reaches into brand.co's own client accounts and pulls their top creative with the performance behind it. Both live in the same platform, so the category read and the read on their own work land in the same report instead of arriving from two separate research efforts.
brand.co also runs its own AI creative system, built in-house, that drafts reports and pressure-tests strategy. Systems like that can't watch video. Structured text is the only form in which a category can be read by software at all, which is why Upspring ended up inside that system rather than beside it. "My AI system does the data work. I do the thinking," is how Casey puts it.
A shorter list of decisions, evidence attached.
Upspring processes each ad into attributes: hook, claim, offer, format, and the page behind the click. Reading text is far cheaper than watching video. Processed creative costs roughly a tenth of what it takes to push raw ads and landing pages through a model, which is the difference between covering a category and sampling it.
That changed one campaign before it started. A draft strategy for one brand rested on a confident claim: the competitor's landing pages were the weak link, so the opening should go after the click-through experience. Upspring's ad-to-landing-page data across that competitor's live ads said otherwise. The pages held up against the ads pointing at them, and the claim died before it reached a client. The thesis that replaced it was different and stronger. Being wrong on a Tuesday costs nothing. Being wrong in a shipped campaign costs a quarter.
What comes out the other end is a shorter list of decisions with the evidence for each one attached.
Strategy that scales with the portfolio.
Research no longer grows in step with the brand count, so adding a client does not mean adding a week of manual scanning.
Full category reads that meant watching ads one at a time now come back as text, per brand, in minutes.
Upspring sits inside brand.co's standing reporting cycle, not a one-off research project.
Coverage stops trading against depth, so judgment scales with the roster instead of thinning across it.
The judgment stays with the strategist. What changed is how much of the category that judgment gets to stand on.