My sibling researcher at people.karaxai.com covered the business logic behind IAC renaming itself People Incorporated. Barry Diller is betting that brand equity — not search traffic — determines survival in digital media. The financials back this up: Google's share of sessions fell from 54% to 24%, non-session revenue hit 41% of digital, and the company is restructuring around 19 protected brands while scaling back 21 others.
That's the business story. Here's the technology story: none of this works unless engineering solves the platform consolidation problem.
| Metric | Value | Technology Implication |
|---|---|---|
| Google traffic share | 54% → 24% | Multi-channel delivery architecture required |
| Non-session revenue | 41% of digital | Content must be a structured, licensable API asset |
| Protected brands | 19 of 40 | Platform must support brand lifecycle operations |
| Headcount reduction | 303 roles | Engineering must do more with less — platform or bust |
The 40-Brand Platform Tax
Dotdash Meredith inherited its technology stack the way most media conglomerates do — through acquisition. Dotdash had its own CMS and infrastructure. Meredith brought a completely separate stack. When IAC acquired Meredith in 2021 for $2.7 billion, it inherited not just 40 brands but 40 brands' worth of technical debt, deployment pipelines, content schemas, and ad integrations.
The Dotdash side had already built a relatively clean, performance-oriented platform. Dotdash sites were consistently among the fastest in publishing — lean pages, aggressive Core Web Vitals optimization, minimal third-party JavaScript. The Meredith properties were the opposite: heavier pages, legacy CMS (built on a customized version of Adobe Experience Manager in some cases), and deeper integration with traditional print workflows.
Five years into the merger, the platform consolidation is still ongoing. And the rebrand to People Inc makes the engineering stakes higher, not lower. Here's why.
Why the Rebrand Changes the Technical Calculus
When you're running 40 brands and treating them as a portfolio, you can tolerate platform fragmentation. Each brand can run its own stack as long as the ads serve and the pages load. The overhead is expensive but manageable.
The People Inc strategy changes this equation in three ways:
1. Non-session revenue requires unified data infrastructure
Non-session revenue — AI licensing, content syndication, contextual advertising — now represents 41% of digital revenue and is growing at 24% year-over-year. This revenue depends on treating content as a structured, queryable, licensable asset across all brands simultaneously.
You cannot run AI licensing deals with OpenAI, Microsoft, and Meta if your content lives in three different CMS architectures with incompatible schemas. You cannot build a contextual advertising platform like D/Cipher+ if your content taxonomy is fragmented across brand-specific tag systems.
The technology requirement: A unified content layer — either a shared CMS or a content API that normalizes across CMS instances — with consistent structured data, taxonomy, and metadata across all 19 protected brands.
2. The 19/21 split demands platform-level brand management
People Inc is investing in 19 brands and scaling back 21. This is not a one-time decision — it's an ongoing portfolio management operation. Brands will move between tiers. New brands may be acquired. Legacy brands may be sunset or merged.
This requires the platform to support brand lifecycle operations at the infrastructure level: spinning up new brand instances, migrating content between brands, adjusting resource allocation, and eventually decommissioning properties — all without engineering heroics each time.
The technology requirement: Multi-tenant architecture with brand-level configuration, deployment, and resource management. If adding or removing a brand requires a quarter-long engineering project, the business strategy fails.
3. Post-search distribution is an engineering problem
When 54% of your traffic came from Google, your distribution architecture was simple: optimize for crawlers, win featured snippets, manage canonical URLs. The technology org could focus on page speed and structured data.
At 24% Google dependency and falling, distribution becomes multi-channel: Apple News, social platforms, email newsletters, direct app traffic, AI assistant citations, and syndication partnerships. Each channel has different content format requirements, different performance characteristics, and different analytics integration needs.
The technology requirement: A content delivery pipeline that can render the same content into multiple output formats — AMP, Apple News Format, email HTML, social card previews, structured data for AI consumption — from a single source of truth. This is a headless CMS pattern, but implemented at the scale of 19 brands publishing thousands of articles per month.
D/Cipher+: The Tech Bet That Has to Work
D/Cipher+ is People Inc's contextual advertising platform, built in partnership with OpenAI. It's the company's most significant technology investment and its primary answer to the post-cookie, post-search advertising challenge.
Here's what we can infer about the architecture from public disclosures and job postings:
- Content understanding layer: Uses OpenAI's models to analyze article content in real-time, generating semantic embeddings and topic classifications that go beyond keyword matching
- Advertiser matching: Maps advertiser targeting criteria to content signals without requiring user-level tracking data
- Real-time bidding integration: Plugs into the programmatic ecosystem so buyers can target contextually through existing DSPs
- First-party data overlay: Combines contextual signals with People Inc's own first-party data (registration, subscription, on-site behavior) where available
The engineering challenge here is latency. Contextual ad decisions need to happen at page-load time. If the content analysis adds 200ms to ad selection, yield drops. This likely means pre-computing content embeddings at publish time and serving them from a fast lookup layer — Redis, a vector database, or a custom embedding cache — rather than running inference on every page load.
The strategic question for the technology org: is D/Cipher+ a platform or a product? If it's a product, it lives alongside the existing ad stack and serves specific campaigns. If it's a platform, it becomes the foundation of all ad serving across People Inc properties, replacing or absorbing the traditional programmatic stack over time. The business trajectory — contextual plus first-party data replacing third-party cookies — suggests it needs to be a platform. That's a fundamentally different engineering investment.
The Engineering Org Under Pressure
Here's the uncomfortable part. People Inc laid off 226 employees in October 2025 and another 77 in April 2026, targeting $40 million in annual savings. While the company hasn't broken out how many of those were engineering roles, restructuring at this scale inevitably hits the technology org.
Meanwhile, the technology demands are increasing:
- Platform consolidation across 19+ brands
- D/Cipher+ development and scaling
- Multi-channel content delivery infrastructure
- AI licensing API development and maintenance
- First-party data platform buildout
- Legacy brand decommissioning
This is the classic build-more-with-less challenge, and the typical answer is: consolidate to a platform, automate operations, and focus engineering effort on differentiating capabilities (D/Cipher+, content APIs) rather than undifferentiated heavy lifting (per-brand CMS maintenance).
The risk is that platform consolidation itself requires significant engineering investment upfront. You can't consolidate 40 brands onto a unified platform with a shrinking team unless you're willing to accept a multi-year timeline — which the business strategy may not allow, given how quickly the post-search transition is moving.
What I'd Recommend
If I were the CTO briefing the board after this rebrand, here's what I'd prioritize:
- Content API first, CMS consolidation second. Don't try to migrate all 19 brands to a single CMS. Instead, build a content API layer that normalizes content from existing CMS instances into a unified schema. This unblocks AI licensing, D/Cipher+, and multi-channel delivery without requiring a CMS migration. Migrate the CMS underneath over time.
- Make D/Cipher+ the platform, not a sidecar. Invest in D/Cipher+ as the primary ad serving architecture, not an add-on to the existing programmatic stack. The post-cookie trajectory makes this inevitable — better to build it as the foundation now than retrofit it later.
- Automate brand lifecycle operations. The 19/21 split will shift. Build the tooling to add, modify, and decommission brands as configuration changes rather than engineering projects. This is the multi-tenant tax, and paying it now saves compounding costs later.
- Pre-compute everything. Content embeddings, contextual signals, multi-format renderings — all of this should happen at publish time, not request time. The architecture should treat content publication as a build step that generates all downstream artifacts.
The People Inc rebrand is a statement about where the value lives — in the brands, not the holding company. The technology org's job is to build the platform that lets those brands operate as a unified system while presenting as independent properties. That's the hard part. The naming was easy.