Product leadership at scale
Pixability
Scaling a pipeline to deliver a $2M annual contract
A nearly full pipeline scaled to deliver a $2M annual contract, lowering operating costs and enabling revenue recognition a quarter ahead of plan.
Service: Product Strategy & Go-to-Market
$2M
Annual contract value
25M
Videos harvested daily
~100M
ML predictions processed daily
~25%
Estimated reduction in pipeline operating costs
Selected work from my time as VP of Product at Pixability (2020–2024).
The problem
Pixability had signed a data-licensing deal worth $2M annually with a programmatic media partner that valued its YouTube insight labeling. Delivery of new API capabilities and a major pipeline scale-up before year-end was an explicit condition of the deal.
We committed to the delivery around September. With the holidays approaching, we needed to finish by early December. The pipeline was already nearly at capacity, and we needed to increase throughput while reducing its operating costs. There was also no API for the channel scoring and channel-library delivery the partner needed to sell contextual advertising to its clients.
Define the delivery
Partner requirements
- API inputs and outputs
- Real sample data
Make room to deliver
Pipeline capacity
- 25M videos harvested daily
- ~25% lower operating costs
Weekly partner check-ins resolved questions.
What I did
Define the delivery with the partner
I authored and refined the specification with the partner, modeling the API's inputs and outputs against real sample data before engineering started.
I secured executive approval and set the delivery plan. A product manager on my team ran daily execution; engineering built the API. Weekly partner check-ins let us test assumptions and resolve questions throughout the work.
Find capacity in the existing pipeline
I evaluated capacity and cost options with engineering. As part of simplifying the pipeline, we removed an in-memory lookup used to check videos and channels against harvesting rules.
The pipeline reached 25 million videos harvested daily. At roughly four ML predictions per video, that supported about 100 million predictions per day. Pipeline operating costs fell by approximately 25% at the higher capacity.
The outcome
We delivered in time to meet the condition the $2M annual contract depended on. The partner described it as the smoothest integration they had done.
Delivery enabled revenue recognition in Q4, ahead of the original Q1 plan. The pipeline could handle the partner's requirements at a lower operating cost.