Ask most engineering orgs about SEO, analytics, or anything that happens before login and you'll get variations of the same answer: "That's marketing." It isn't. It's the top of your product. And the habit of treating it as someone else's problem quietly taxes acquisition, activation, and roadmap quality every week.

How We Got Here

We built habits in a different internet. Client-side frameworks made CSR the default, and "we'll fix SEO later" became culture. Then privacy and browser changes softened client-side analytics to mush. Meanwhile, agencies and in-house teams optimized for velocity over visibility: ship features, show demos, move tickets. The result: beautiful SPAs with brittle discoverability, dashboards that don't connect to revenue, and a roadmap funded on vibes.

The technical shifts were predictable but ignored. JS-first everything meant crawlers could theoretically render your content, but it's fragile and slower than giving them usable HTML. Google now treats dynamic rendering as a last-resort workaround, not a strategy. Page experience became a ranking reality with Core Web Vitals as first-class quality signals. Ignoring them hurts both search and conversion. Client-side tracking broke as Safari's ITP and related protections shortened cookies and killed legacy patterns. First-party and server-side approaches became the only stable path forward.

None of that is a marketing problem. Every one is an engineering choice that engineering pretended wasn't their responsibility.

The Cost of Pretending "That's Not Our Job"

When you treat growth as someone else's problem, you pay compound interest on that decision forever.

You give up free distribution. CSR routing without SSR, SSG, or prerendering makes indexing harder. Your sitemaps go stale because nobody owns them. Your structured data never lands because "that's SEO stuff." You ship features nobody discovers because Google can't figure out what you built. You're essentially building a store in the desert and wondering why foot traffic is low.

You lose evidence about what actually works. Front-end beacons undercount by 30-40%. Ad-blockers strip your tags. Cookie lifetimes collapse to nothing. Your "data-driven" team makes decisions based on partial data that skews toward technical users who don't block trackers. You're not data-driven; you're flying on instruments that lie.

You outsource your narrative to randomness. Without Organization schema and canonical MCPs (Minimum Credible Pieces), AI assistants and SERPs guess who you are based on whatever they can scrape. They often guess wrong, pulling from competitor blog posts about why you suck or forum threads from three years ago. You've let machines write your elevator pitch, and they're terrible at it.

The Real Reasons Engineers Dodge This

Engineers avoid growth work for reasons that made sense in 2010 but are dangerously wrong now.

"SEO is just keywords" was never true, but now it's obviously false. SEO is rendering strategies, routing decisions, performance budgets, and structured data. It's deciding whether your framework can produce HTML that search engines can read. It's choosing URL structures that won't break when you refactor. It's pure engineering.

"Analytics is marketing's tag manager" assumes tags still work. Under ITP and modern privacy controls, client-side tags are basically suggestions. Real analytics requires server-side capture, first-party domains, and event schemas designed into your API. Marketing can't fix that with Google Tag Manager.

"Customer interactions equal support tickets" ignores that docs, changelogs, Google Business Profile, and reviews are upstream product surfaces. They shape perception before users ever hit your landing page. They're product decisions implemented in code and content systems.

"We don't have time" is the most expensive lie. You're already paying in lost organic traffic that compounds monthly. You're paying in weak attribution that makes every growth experiment questionable. You're paying in slower learning because you can't trust your data. The time you "save" by ignoring this costs you multiples in engineering effort spent on the wrong things.

The Minimum Viable Growth Stack

Stop treating these as nice-to-haves and implement them as platform requirements.

Render for Humans and Machines

SSR or SSG for primary routes isn't a performance optimization; it's table stakes for discoverability. Stable, descriptive URLs that don't break when you ship. Canonical tags that consolidate ranking signals. HTML titles and descriptions set server-side, not injected by JavaScript after the page loads. Internal links that work without JavaScript. XML sitemaps that reflect actual inventory and update automatically.

Google's own guidance is clear: make JavaScript search-friendly, build sitemaps search engines can trust. This isn't advanced; it's baseline. Yet most SPAs fail these basics because nobody put them in the requirements.

Make Your Entity Legible

Add Organization structured data with robust sameAs links pointing to your site, LinkedIn, GitHub, and review pages. This isn't just SEO; it's teaching every machine on the internet who you are. It stabilizes what shows up in knowledge panels, AI overviews, and assistant responses.

Without this, you're hoping search engines and AI models correctly infer your identity from context. They won't. They'll confuse you with competitors, merge you with similarly named companies, or simply mark you as "unclear" and show nothing.

Instrument Reality, Server-Side First

Log activation and revenue events from your API, not just window events that get blocked. Route client SDKs through a first-party subdomain to reduce data loss. Stop pretending that Google Analytics default setup gives you real data. It doesn't.

Server-side and first-party setups aren't just more accurate; they're the only thing that works reliably now. Safari caps client-set cookies at 7 days. Brave blocks most tracking by default. Even Chrome users increasingly run blockers. If your analytics lives in JavaScript, you're measuring a biased subset of technical users.

Budget for Speed

Treat Core Web Vitals like uptime. Set performance budgets in CI that fail builds when violated. Run synthetic checks that alert on regression. Collect field telemetry that shows real user experience.

This isn't just about SEO rankings anymore. Users bounce from slow sites regardless of how they found you. Every 100ms of latency costs conversions. Speed compounds: faster sites rank better, convert better, and get shared more.

Docs as an Entry Point

A 15-minute quickstart that actually works. Diátaxis structure (tutorial, how-to, reference, explanation) that serves different user needs. Search that doesn't make users guess the exact term you used.

Documentation isn't support cost; it's user acquisition. Stripe and Twilio built billion-dollar businesses on great docs that made developers feel competent. Your docs are your product for many users. Treat them accordingly.

Make It Routine: The Discoverability & Evidence Checklist

Replace "remember to do SEO later" with a checklist you run every sprint, embedded directly in your PRD template.

Rendering: Route is SSR/SSG or pre-rendered. Titles and meta set server-side. Canonical URL configured. No JavaScript-gated content that search engines need.

Structure: Organization JSON-LD updated. Relevant page-level schema added. Internal links work without JavaScript. XML sitemaps automatically updated.

Experience: Core Web Vitals budget defined (LCP under 2.5s, INP under 200ms, CLS under 0.1). Regressions blocked in CI. Performance monitoring in production.

Evidence: Activation and revenue events emitted server-side. First-party or server-side tagging configured. Dashboards updated with new events.

Make this as routine as code review or security sign-off. It's not extra work; it's the work that makes the rest of your work discoverable and measurable.

Who Should Own This

You need a growth-minded product engineer who understands AARRR funnels and can implement them. Someone who writes production code across the stack, not just frontend or backend. Someone comfortable with feature flags and experiments as part of normal development. Someone who can discuss rendering strategies and privacy trade-offs without needing a translator.

This person sits in engineering, not marketing. They pair with PM and design like any product engineer. They treat discoverability and evidence as platform concerns that affect every feature. They're not doing "marketing work"; they're doing engineering work that happens to drive growth.

Without this role, you're back to hoping marketing can fix rendering problems with meta tags. They can't. You're hoping agencies can patch your analytics with tag manager. They can't. You're hoping growth happens accidentally. It won't.

The Agency Problem

Even top agencies miss this because incentives push shipping screens, not shipping surfaces. Award-winning SPAs get built with no SSR or SSG, no schema, 404s on deep links, sitemaps that don't update, and analytics stitched together with brittle tags. Marketing teams inherit these technical disasters and do their best with CMSs and tag managers, but they can't patch rendering, routing, or event capture from outside the codebase.

Agencies bill for visible deliverables: designs, pages, features. They don't bill for "made it indexable" or "instrumented activation correctly" because clients don't know to ask. The result is beautiful products that nobody can find and nobody can measure.

The Bottom Line

If your org still treats SEO, analytics, and upstream customer interactions as "marketing things," you're leaving compound growth on the table. Not capturing it once means not capturing it forever because growth compounds.

This isn't about blog posts or social media or any traditional marketing. It's about rendering, routing, structure, speed, and evidence. These are engineering decisions that determine whether your product can be discovered, whether you can measure what works, and whether machines represent you accurately.

Every week you postpone this is a week where competitors who get it pull further ahead. They're not smarter; they just stopped pretending growth was someone else's job.

Work with withSeismic

This is exactly what I build: growth infrastructure that lives in engineering where it belongs. We implement the Minimum Viable Growth Stack, embed discoverability and evidence into your development process, and build the dashboards that prove impact. No hoping marketing can fix it later. No agency band-aids. Just engineering solutions to engineering problems that happen to drive growth.

withSeismic (Doug Silkstone). 6-week Growth Engineering Sprint. Code shipped. Growth owned. Engineering solved.