From Checkbox Enablement to Sub-30-Day POCs at Treasure Data

AUTHOR: Michele Nieberding

Treasure Data is a customer data platform that unifies online and in-store customer activity so brands can market more relevantly. I led product marketing and sales enablement there while we layered AI onto the CDP and pivoted messaging toward AI products and suites. We already had a standing weekly rhythm: one dedicated hour of sales enablement every week plus a separate hour of product office hours the following day. Enablement decks and topics were produced on a regular cadence and recordings were shared, which looked good on paper for a fast-moving MarTech company going through what felt like an identity crisis in the AI category.

The cadence had turned into a hamster wheel. Topics got lost, decks were thrown together at the last minute, sales often skipped live sessions in favor of recordings, and there were no real feedback loops on what was working. We weren’t tying enablement to outcomes—revenue, pipeline, sales-cycle length, or win rates—and we were well below revenue goals. Deal analysis showed two clear funnel gaps: weak top-of-funnel hooks that weren’t generating enough qualified enterprise pipeline, and late-stage losses to a major competitor. Reps were drinking from a fire hose of weekly topics with almost no reinforcement, so recall and retell ability suffered. Simply “fixing messaging” would have been the easy (and wrong) diagnosis.

I went back to basics. First I diagnosed the signals—missed revenue targets, pipeline volume, stage-by-stage drop-offs, and the spread between top reps and struggling reps—using win-loss data, Gong calls, and CRM. Then I talked directly to both high performers and lagging reps with carefully framed questions (“What deals keep you up at night?”) so they wouldn’t go on defense. That produced a clear hypothesis: the live sessions themselves weren’t the main problem; the missing piece was reinforcement and a tight feedback loop so reps could actually articulate the new AI story. From there we ran a structured experimentation program: internal brainstorms and mini-hackathons to surface formats, then rapid tests of human-led and AI-assisted methods, always measuring engagement, recall, and whether the content changed real conversations. Successful tactics were baked into the ongoing enablement rhythm; weak ones were killed. We also extended the work into post-call follow-up assets and kept scoring what moved pipeline and cycle time so we could take proof back to leadership.