SEO for publishers is a different job from SEO for a business with forty pages. A publisher has an archive that grows every day, a newsroom that will not take instructions from a marketing consultant, an advertising stack that fights the performance budget, and traffic that arrives in bursts from surfaces nobody controls. The levers that matter are structural: what gets crawled, what stays indexed, how quickly a page renders with ads on it, and how well the site describes its content to machines. This page sets out where publisher search work actually pays, and how to judge a provider who claims to do it.
The archive and crawl problem
At publisher scale, the binding constraint is usually not whether a page is good but whether it is crawled, indexed and still discoverable a year later. Google's documentation on managing crawl budget for large sites explains that crawling is governed by capacity limits and by demand, and that a slow server or a large volume of low value URLs consumes crawl capacity that could go to content you care about. Publishers generate low value URLs prolifically: tag pages, paginated archives, author pages, session parameters, print variants, syndicated duplicates. The work is a mixture of pruning, canonicalisation and internal linking so that the pages worth keeping remain reachable in a few clicks from something that gets crawled often. The second half of the problem is decay. Old articles lose their internal links as they fall off the homepage and section fronts, so the archive quietly becomes unreachable while remaining published. A competent provider will have a systematic approach to this, usually some combination of evergreen hubs, related content modules built on real relevance rather than recency, and a defensible policy on updating versus retiring old material. Ask for that policy explicitly, because it is the difference between an archive that compounds and one that only costs hosting.
Surfaces, structured data and what they are worth
Publishers earn traffic from several distinct surfaces and they behave differently. Standard search results reward depth, coverage and authority over time. Google Discover, which Google documents as a feed of content tailored to user interests, is driven by interest rather than by a query, which makes it volatile and unusually sensitive to imagery and headline framing. Google's own Discover documentation notes that large, high quality images help content appear there, and that content can appear without any specific markup requirement beyond the usual policies. Structured data is the connective tissue: Google's article structured data documentation describes the properties that help search understand a news, blog or sports article page, including headline and image information. None of these surfaces should be treated as a strategy on their own, because dependence on a volatile feed is how publishers end up with a traffic cliff and no plan. The sensible posture is to build for the search surface you can influence steadily, take the Discover upside when it comes, and keep a direct audience relationship, meaning newsletters and apps, that no surface change can remove. Ask a candidate provider how they would reduce your dependence on any single source rather than how they would maximise one.
Where advertising and performance collide
Publisher sites are slow for a reason: the advertising stack pays for the journalism. That creates the central tension of publisher technical work, because header bidding, consent tooling, video players and analytics tags each add weight, and page experience signals are documented by Google as part of how it assesses pages. The resolution is never to strip the ad stack. It is to negotiate it: lazy loading below the fold, tightening the number of demand partners, reserving space so layout does not shift, and measuring revenue against speed rather than assuming either wins. A provider who proposes performance improvements without asking what each tag earns has not understood the business they are optimising. The commercial layer also has legal edges. Affiliate content is now a substantial revenue line for many publishers, and the Federal Trade Commission's endorsement guidance requires that material connections between an endorser and a seller be disclosed clearly, which for a publisher means visible affiliate disclosure on commerce content rather than a line in the site footer. Ask any provider how they handle disclosure placement on commerce content, because getting it wrong is a regulatory risk, not a design preference.
Vetting a publisher SEO provider
Ask for a named publisher client and the scale they worked at, since the practices that work on a site with a few thousand articles do not all transfer to one with hundreds of thousands. Ask how they work with editorial, specifically whether they train journalists, sit in news meetings, or file tickets to a development team, because a provider who cannot get anything adopted in a newsroom will deliver excellent advice and no change. Ask what they would do in the first month, and be suspicious of an answer that begins with keyword research rather than with logs, indexation and the archive. Ask how they measure, and insist that the reporting distinguishes surfaces, since a stable search number can hide a collapsing feed number and vice versa. Ask who does the technical implementation, because most publisher gains require engineering time that an agency cannot supply. Many publishers buy this as part of broader digital marketing and SEO services; if you do, insist that the publisher specific work is scoped and reported separately, since it is the part that behaves least like ordinary marketing and is the first to be quietly deprioritised.
Questions people ask about seo for publishers
Should we delete old articles?
Rarely delete, frequently consolidate or update. Articles with inbound links and lasting relevance should be maintained and relinked. Genuinely obsolete material that attracts nothing can be retired with redirects to the closest useful page. A blanket pruning policy applied without checking links and traffic is how publishers lose value they cannot get back.
How much does site speed matter with a full ad stack?
It matters, and it is negotiable rather than binary. Google documents page experience signals as part of how pages are assessed, so the practical work is reducing weight where it costs the least revenue: deferring below the fold units, reserving layout space, and trimming demand partners that contribute little. Measure both sides before cutting.
Is Google Discover worth building a strategy around?
Worth building for, not depending on. Strong imagery, honest headlines and content people want to read give you a chance at it, but the feed is interest driven and can change without warning. Treat it as upside on top of a search and direct audience base rather than as the base itself.
Do we need a specialist or will a general SEO agency do?
Scale and newsroom dynamics favour a specialist. The problems are crawl management, archive maintenance, structured data at volume and negotiating with an ad stack, and general agencies rarely have that experience. Judge on a named publisher client at comparable scale rather than on the label.