How mongodb.com moved from ad hoc requests to a single intake, tiered service levels, a documented workflow for every kind of work, and commitments that run both ways between the web team and the teams it serves.
The problem
In 2022, MongoDB's web strategy called for retooling the marketing site and building a new way to run it. Requests for mongodb.com arrived through meetings, email, and Slack, with no shared intake and no common definition of done. Product marketers were building pages themselves, turnaround depended on who you asked, and once a page launched, nobody clearly owned keeping it current.
The approach
The goal was a model any team could ship through without renegotiating the process each time. It rests on five working principles.
- One front door. Every request goes to a web producer, who scopes it, sets the web strategy for it, and owns it through launch. A request can start as a Slack thread, but it becomes a ticket before work begins, so every piece of work has a record, an owner, and a date.
- Size the work before starting it. Each request is sorted into one of three tiers, from a quick fix to a net-new launch. Each tier carries a published turnaround and a checklist of what the requester brings. Turnaround depends on materials arriving on time, so the checklist works as the contract.
- A documented path for every kind of work. New pages, page updates, events, lead pages, press releases, redirects, and urgent updates each follow their own written workflow. Routine work goes straight to publishing, which keeps producers on the work that needs strategy.
- Commitments run both ways. Requesting teams bring goals, success metrics, go-live dates, and visibility into their roadmaps. The web team brings proactive updates, recommendations grounded in data, and reporting on how each page performs.
- Quality gates, then close the loop. Every role in the workstream signs off on the copy before anything reaches production. After launch, standing reviews with each partner team keep pages current through a continuous cycle of review, measure, and update.
Where work repeats, it is automated. AI writing assistants draft copy for simple requests, with a copywriter reviewing for brand and tone, and webinar requests now set up their own forms, campaigns, and tickets as soon as they are approved.
My role
I co-established the operating model with my peers on the Web Experiences team, building on a web strategy the team defined together. The day-to-day detail is mine: I built the service-level framework, the requirements for each activity, and every workflow the model documents.
The result
+34%, then +75%Growth in published output, year over year, with no added headcount.
Every launch date held through a multi-year CMS migration and full site redesign, and every team that publishes to mongodb.com now works through the model. The product page rollout is one of the programs that runs on it.