Infrastructure enables personalization

People Inc. is “America's largest publisher”—the consolidated entity formed when Dotdash and Meredith merged under IAC, since renamed from Dotdash Meredith. During my tenure from early 2024 to late 2025, the network spanned 40+ brands and 22M+ users, with editorial properties ranging from People and Entertainment Weekly to Travel + Leisure, Food & Wine, Allrecipes, and Investopedia, to name a few. I was hired into the product organization as senior PM for audience relationships. The written mandate was newsletter migration work: moving the network's newsletter and sweepstakes experiences onto a shared MarTech stack as part of the post-merger consolidation. The implicit mandate was that I was being trusted to develop and drive the broader growth vision—and the early days of the role were spent pitching that vision consistently until it became the actual shape of the work.

The platform I walked into shaped the work that followed. The Overlay was the foundational onsite marketing surface across the network—an engagement-triggered modal that served a generic newsletter signup offer—and it ran on Digioh, the marketing tool that read the rendered DOM for targeting. Acquisition was per-newsletter, with no concept of a user account underneath; the CDP existed alongside it, surfacing behavioral patterns into analytics. The marketing organization owned a roadmap of newsletter launches queued against the migration, and that roadmap was the explicit deliverable I was hired against. Executing the vision, however, required a broader product roadmap.

I co-owned growth programs with marketing and product leadership; that was my primary team. I worked daily with my VP, a VP of audience development, and several marketing directors who each owned a part of the lifecycle: conversion, retention, operations, and platform. My product team included a tech lead, manual and automated QA, and three engineers, plus access to a platform designer on a project-by-project basis (I mostly delegated user research to her and mentored her as she refined her skills in this area). Every other partner—data science, data operations, design operations, brand product teams, editorial leadership, privacy, and legal (again, to name a few)—sat fully on another team and contributed to my team's work as a result of my influence. Roughly halfway through my tenure, a restructuring shifted my reporting line, reset the team's resourcing, and reinforced the commitment to my vision for audience and engagement growth. Between Q4 2024 and Q1 2025 the role moved from executing a backlog of marketing projects to driving a multi-year roadmap for identity, registration, and onboarding.

The institutional rallying cry was “jumping the wall”—the bet that the brands' sites should become destinations on their own terms, with directly-owned relationships that didn't depend on Google referrals or social platforms surfacing the content. SEO and the emerging GEO (generative engine optimization) layer still mattered, but newsletters carried a different kind of weight in this thesis: the inbox was framed as a rarefied, sacred space—a directly-owned channel to the user that the platforms in between couldn't intermediate. Niche, topical content and personalization were the long-arc vision; the organization was bullish on both.

What leadership couldn't see was the infrastructure that vision required. The CDP's behavioral signals lived in analytics; they never reached the product or the newsletter experience in a way that could shape what a given user saw next. The theory of change—personalize the site experience, drive time on site, lift ad revenue—was correct. The gap was that modern personalization runs on user-account infrastructure and user-controlled account management, and the network's platform had neither. The vision was right; the infrastructure shape underneath it didn't exist.

The reality on the user side made the gap concrete. A subscriber to the People News newsletter received 20+ editorialized emails a day, with no underlying capability to personalize what arrived when. On site, the newsletter signup units that popped up never acknowledged whether the user was already subscribed; you could be offered a newsletter you'd signed up for the week before, and the signup flow itself wouldn't confirm what you already had. The qualitative truth was widely known across the organization. What was missing was a clear, quant-driven roadmap of solutions that leadership could buy into.

An additional gap sat underneath, almost undetected: marketing capacity. I surfaced it continuously, and together with the identity infrastructure gap it became the opportunity space I delivered within. My work was not only about delivering surfaces, capabilities, and programs, but also about doing so in a way that alleviated marketing capacity for higher-leverage activities.

Two strands of evidence converged on the diagnostic. One I translated; one I built.

The third piece of discovery wasn't about the user experience or the platform—it was about how the organization itself shipped. Across every initiative, I documented intake, handoff, and delivery acceptance criteria, with PRDs nested under larger initiative milestones and a RACI chart agreed upon at kickoff for each initiative. As points of failure accumulated, the documentation made bottlenecks legible. Systems thinking and information architecture are superpowers of mine that showed up frequently in how I executed my role: organized product documentation, a modular roadmap with clear dependencies, and project plans with clear roles and timelines. This rigor around artifact creation and maintenance is also the foundation of AI-native operations.

The bet was co-shipping. My team wasn't plugging marketing programs into existing capabilities and pre-built surfaces; it was building the surfaces and capabilities at the same time as the growth programs that proved them out. Each program landed when its capability landed, on the same release cadence, in a tight but modular sequence: this program needs this surface and this capability, so build them in lockstep. I stewarded the longer-arc bet underneath that loop—the identity, registration, and onboarding infrastructure underpinning every program—in the background of the marketing organization's newsletter-program-launch roadmap, advanced one increment of buy-in at a time as each program proved the next part of the thesis.

The architectural shape that made network-scale rollout feasible was a three-tier enablement pattern. The first tier decided which pages the Digioh script loaded on. The second tier decided which pages the empty div placements were added to—each unit type had its own implementation, but every one of them spoke the same empty-div-then-unit-injection contract so that the marketing team could iterate and ship new program creative for any units without requiring engineering resources. The third tier was the per-campaign targeting tracked in a visible campaign tracker, with marketing controlling targeting and creative inside Digioh and email content inside Iterable. Brand teams had the option to inherit the network defaults at each tier and could override them where their own roadmap or audience shape demanded it. That core-inheritance-with-brand-override pattern was a lesson learned the hard way (more on that below); once we adopted it, every subsequent unit shipped faster and survived handoff to the brands.

Day-to-day, sequencing the work involved three factors: team capacity, roadmap dependencies, and revenue modeling. Technically, capability and surface dependencies governed which program could ship when. A brand wanting to prioritize a program that relied on the new capability and surface had to prioritize enabling that capability and surface first. To that end, per-brand rollout was governed by each brand team's own roadmap and capacity as much as by the implementation's revenue impact on each brand. I collaborated with marketing to develop the revenue models for each program, then built and pitched each program's rollout plan, taking into account the model projections and the constraints of each brand's roadmap. I gauged stakeholder interest and buy-in early, usually off of a basic revenue model that we'd iterate on as we got data from brand launches. Throughout the process, we also had to meet brand expectations regarding impacts on site performance, both as a dimension in its own right and as a downstream lever on revenue. As we got closer to release, I ran regular demos, office hours, and similar touchpoints to build alignment, gather any additional feedback, and secure buy-in on the rollout plan. My team's capacity, including marketing, was the third sequencing factor.

For longer-term planning, I also mined other teams' experiments; I was looking out for two things: free intel and possible work that could serve as a proxy for work my team wanted to do. In the latter case, we could leverage their data to build early models that would help us sequence certain work sooner (or even bypass it entirely). For example, when the People team announced they'd be testing login in the header for comments, I regularly checked in on their work and data to move site registration up in my team's sequence with increased confidence and lower predicted effort.

What follows is a 0-to-1 across surfaces and capabilities, then a 1-to-ninside a single compounding program—together, the meta-0-to-1 the role was sized against.

Every capability, surface, and program followed the same shape: 0 to 1 at a pilot brand, then 1 to n across the network. The 0-to-1 half was capability invention—surfaces, data plumbing, governance patterns—done once. The 1-to-n half was sequencing—brand roadmaps, capacity, revenue models, demo cadence, override patterns—repeated per brand. Both halves had to land for a launch to count, and the work that survived handoff was the work that treated the second half as first-class from day one rather than as a rollout afterthought. I'd internalized the discipline at User Interviews, where shipping into a two-sided marketplace meant every supply-side change had to land against demand-side recruitment running in parallel.

What follows is the 0-to-1 foundation in four pieces: capability prerequisites, the inline-units misstep that hardened the brand-override pattern, the toaster + slide-in pivot that surfaced the unit-design discipline, and Comments—the program that proved the identity bet had legs and seeded the email-journey work Follow This Topic (below) would lean on.

Follow This Topic was the program built to leverage the 0-to-1 work for personalization. The surfaces, the data plumbing, the brand-override pattern, and the revenue models were tests of whether the platform could carry a set of relatively simple programs from pilot to network rollout. This was the test of whether it could handle a program with far more complexity than the others.

Follow This Topic was the topical-affinity bet—the “personalized” newsletter program and the launch the rest of the platform was built for. The hypothesis from our user research was that readers wanted topical newsletter subscriptions surfaced at two moments in the article: at the top of the page before they read, then partway through the article body as a reminder once they'd engaged meaningfully with the content. We built the MVP on Travel + Leisure. The Digioh script was already on the majority of the site from the platform expansion; we placed a feature-specific empty button div at the top of a subset of pages and shipped Follow This Topic with a future plan to test into the inline unit addition, both placements triggering the Overlay with the topical signup offer.

Page selection was modelled off four factors:

• Traffic to pages by taxonomy
• Publication rate of new content per taxonomy
• A simplified brand newsletter signup conversion rate
• A simplified brand RPM

The model clarified which topics would drive enough impact to justify the placement. However, it required a load-bearing risk-mitigation move—every topical signup also enrolled the user in the brand's flagship newsletter, so the program protected the existing revenue base while we tested whether a topical channel could pay for itself. I led the partnership between brand marketing and editorial leadership to set the final taxonomy list, the offer and creative combinations, and the email journey that followed—subject lines, sequence, content selection. The retention-side care put us ahead of where we thought we'd be on proving out personalization, and those outcomes became part of our headline in our pitches to brands for post-MVP adoption.

The data-flow work underneath Follow This Topic was just as critical to nail as the user experience.

On the page side: the taxonomy data Digioh needed lived in the source HTML but only sometimes made it to the DOM, per the discovery noted above. We partnered with the SEO team to extract the classification cleanly into the DOM as numeric IDs (the SEO team's concern was that inserting human-readable taxonomy strings would degrade SERP performance through unintended keyword stuffing; numeric IDs solved that without losing the targeting signal). This had a ride-along effect on top-of-funnel volume for all the other programs that had been running starved of clean targeting data since I joined.

On the email side: Iterable received content classification via Hightouch's reverse ETL, and a transformation in that step was producing content mismatches between what should have qualified for a send and what actually did. A sub-initiative inside Follow This Topic re-shaped the Hightouch transformation so the two systems agreed on which content should be associated with which topic. This also had a ride-along effect on the newsletter content relevance, driving greater engagement and return traffic for all the other programs.

The same data work fed two different parts of the marketing lifecycle; the Follow This Topic launch allowed us to broker investment in foundational improvements.

Stepping up a level: this program and the 0-to-1 foundation before it were a meta-0-to-1—proof that the personalization bet could be successfully executed. At this meta level, 1-to-n was the longer-running personalization roadmap the platform had been built to enable, sequenced behind the identity infrastructure stewarded underneath the whole arc.

The network grew email revenue 33% year-over-year during my tenure—against a ~12% YoY baseline the year prior. The growth-program portfolio I owned roughly tripled the network's email-revenue growth rate above baseline. That portfolio spanned capability and surface development, marketing-campaign work, and newsletter-content development across program launches.

The growth-program portfolio I owned roughly tripled the network's email-revenue growth rate during my tenure—33% YoY against a ~12% baseline the year prior. Inside that portfolio, the comments product I architected on People opened a new $2.2M annual revenue channel.

Follow This Topic—the topical-affinity newsletter program I built end-to-end on Travel + Leisure—drove a 3x topical-open rate among signups and 2x customer LTV against the flagship-newsletter average.

The second-order outcome is the rollout playbook itself. Every program in this case study moved through the same loop: build the MVP on one brand, prove it on revenue and site performance, then steward adoption brand-by-brand against each brand's own roadmap and capacity.

People Inc.'s thesis was right—directly-owned user relationships, deeper engagement, ad revenue lift through time on site—but leadership couldn't see that the strategy required a technical foundation that didn't exist. You cannot personalize effectively without identity. You cannot run topical-affinity programs effectively without a data platform that appropriately aligns content classification across each tool that operationalizes the experience. The vision was sound; the infrastructural shape underneath it was the gap; closing that gap was the work. I'd run the pattern before at Muck Rack, where standardizing the content schema was the precondition for everything the platform's search and discovery features built on top of it.

The skills I honed in this role were all about operationalizing vision into strategy and strategy into execution. The first part looked like collaborating with cross-functional teams and leadership to continuously refine the strategy—translating it into the modular sequence of capabilities, surfaces, and programs that drove impact across the line of business. The second part looked like developing artifacts, processes, and systems that supported the execution of the strategy. All of this shaped me into a more strategic product leader who isn't afraid to roll up his sleeves and get his hands dirty. That's the posture I bring to any media company driving personalization bets through platform development.

Two next steps, if this is the kind of PM work you're hiring for: review my resume →.

Or, if you're ready to talk, book a 30-min product chat ↗ (opens in new tab).