Weather-triggered bidding, simulated
The real one ran in Smartly. This one runs in your browser. Randomize the forecast and watch the bid rules react - or tap a city to see the creative follow the weather.
Not to scale. Cartographers, look away. Tap a city to select it.
Sun's out in Phoenix.
Bid up, show the polarized pair.
Bid 1.3xThis is a simplified simulation of a real system I ran in Smartly for a sunglasses client: a weather API feed drove automation rules that set geo bid multipliers and swapped in localized creative. Sunny cities got bid up, gray ones got bid down, which turns out to be a very good way to sell sunglasses. Result: a ~30% ROAS lift.
A raw feed row, dressed for work
A raw catalog feed gives you a white square and a price. A supplementary feed turns that same row into on-brand, market-localized creative at SKU scale. Drag the handle.
One catalog, five markets
Language, currency, seasonal overlays, weather copy, delivery windows, stock levels - all just feed columns. Pick a market or a column below and the same SKU re-dresses itself. No new creative production per region, and it scales to every SKU in the catalog at once.
Localized name, currency, and overlay - straight from the feed.
Multi-region feeds: region-specific supplementary catalog feeds across five markets for a multi-thousand-SKU apparel client drove a ~150% ROAS increase.
BFCM mode: scheduled feed-level promo changes let us scale DPA spend ~5x into the client's most profitable BFCM on record - without resetting the learning phase.
The resume, compressed
Six numbers instead of six paragraphs. Client names withheld out of politeness (and a couple of NDAs).
~7 years agency-side (PHD Worldwide → Reprise → Wpromote → Black Propeller), scaling D2C, B2B, and ecommerce growth on Meta, TikTok, Snapchat, Reddit, Pinterest, etc.