Platform Product Manager

PM Roles — Part 2

What is a platform product management role?

A platform product manager develops horizontal capabilities that can be leveraged across multiple verticals, multiple sites or geographies with support for application specific preferences. Often, this is in the form of delivering a multi-tenant/instance SaaS product that provides leverage across multiple applications, products or brands within an enterprise or serves multiple enterprises. This role isn’t only seen in a platform business, but is commonly seen in all technology companies.

What are some examples of platform products?

Here are a few platform products that I have developed as a product manager : i) Search engine platforms for ecommerce businesses. ii) An Ad platform that is used by multiple publishers to serve sponsored products. iii) An analytics platform that’s used by several banks to serve reports & analytics to the bank’s consumers.

What are sample objectives for PMs in these roles?

The search engine platform served multiple business categories and had category specific ranking features & algorithms. It served multiple sites and languages across many countries. In this role, the key objectives were: i) Reducing time to market for new categories & countries ii) Uptime iii) Maintaining query performance as a function of concurrency iv) Increasing algorithmic efficiency as measured by a search metric such as mean reciprocal rank or NDCG.

In the ad platform, the key objectives were: i) Reduce time to market for new publishers ii) Enable high return on ad spend (RoAS) for advertisers iii) Provide trustworthy sales lift measurement iv) Grow ad revenue processed by the system thereby increasing ad platform revenue.

What are typical ways of working for platform PMs?

Search Engine Platform

The search engine platform at a large eCommerce company had 100s of large categories. I would meet with category PMs, educate them about search APIs, ranking algorithms & understand customer problems. Often, the PMs had domain specific insights which helped to identify failure points in the search engine. Using these qualitative insights, I would develop quantitative methods to come up with sets of problem queries to review with engineering. The discovery process would result in identification of new or updated machine learning features followed by recall, ranking, autocomplete or other algorithmic improvements.

Once new machine learning models were ready, I’d be responsible for testing the models and making launch decisions with engineering, business leadership. A key part of the role was driving alignment with the business teams on which KPIs to optimize. I’d run an A/B test to measure lift in a KPI such as revenue, active users, new subscriptions, # of streams and help to get closure with the business/category PMs to launch the feature.

Ad platform

The ad platform was intended to replace a B2B2C platform provider by bringing related capabilities in-house. Developing this platform required understanding advertiser, publisher requirements & competitive considerations. Post the discovery period, I had several concrete deliverables: i) Ad Server APIs that the publisher’s interface teams would integrate ii) Data feeds and beaconing requirements for the publisher iii) Campaign Management UX requirements iv) Measurement & reporting requirements v) Campaign management APIs for advertisers vi) Data science/algorithm requirements

Once these deliverables were identified, my team published ad server integration documents for front-end developers. Product execution was a co-development effort with these partner teams. I worked with operations teams to develop a launch plan. We went through the standard product execution process which involved sprints, monthly executive updates, a beta launch for selected advertisers, followed by a ‘general availability’ launch. The ad platform will eventually be used by several different publishers across the enterprise delivering cost savings for the publishers, consistent APIs for advertisers, consolidation of supply of ad inventory across the enterprise and developer efficiencies as we continue to develop new ad products.