I recently planned SEO for a sports news and tips site run by a betting news publisher. It publishes a lot: previews, tips, odds round-ups and news, every day, across many sports. The temptation with a site like that is to treat SEO as a volume game. Publish more, target everything, hope something sticks.
The plan I wrote went the other way. It starts by being honest about which searches the site can win, then spends effort there.
SEO for a sports news and tips site starts with search families
Individual keywords are too small to plan around. Families of searches that share an intent are the right size. For a tips site they group roughly like this:
| Family | Example | Winnable? | What wins it |
|---|---|---|---|
| Event tips and predictions | "[team] v [team] prediction" | Yes, briefly | Being early, a named tipster with a visible record, a clear pick and price |
| Competition odds | "[competition] winner odds" | Yes, over a season | One evergreen page at a stable address, kept current |
| Betting terms | "what is each way" | Partly | Worked examples and tools that an answer box cannot give |
| Breaking sports news | "[player] injury" | Rarely | Speed and genuine reporting; general sports outlets usually win |
| Brand and navigational | "[site] tips today" | Already won | Fast, clean pages and good internal links |
Being clear that breaking news is mostly not winnable is useful in itself. It stops the team pouring optimisation effort into pages that were never going to rank, and frees it for the families where the site has an edge.
Evergreen odds pages beat a page per event
A new article for every round of odds splits the authority across dozens of near-identical URLs, each of which goes stale within days. The better pattern is one page per competition or market at an address that does not change, updated as prices move, with the date and age of the prices on the page.
The articles still matter. They link into the evergreen page, and the evergreen page links out to the latest coverage. Over a season the stable page collects the links and the history, and it is the one that ranks.
A glossary needs a quality bar before it gets indexed
Betting glossaries are full of thin entries: a term and a one-line definition. Hundreds of those indexed at once is a good way to tell a search engine the site produces thin pages. The plan set a bar. A term page is indexed only when it has:
- a definition written for a newcomer, not copied from another glossary;
- something an answer box cannot give, such as a worked example, a small calculator, or the price shown in every odds format;
- real articles on the site that use the term, found by matching the term in the text rather than by tagging.
Entries below the bar stay published and useful for readers, but carry noindex, follow until they are improved. That way the glossary can grow without dragging the site's quality signals down, and each improved entry can be promoted on its own.
Structured data that matches what is on the page
Structured data is not a ranking trick. It is a way of stating clearly what the page already shows. For this site that meant:
NewsArticlewith the real author, publication and modification times;Personfor each named author, linking to their profile;DefinedTermandDefinedTermSetfor the glossary;SportsEventon calendar and event pages, with start times only where the time is known;BreadcrumbListeverywhere, matching the visible breadcrumbs.
The rule I hold to is that nothing goes in the markup that a reader cannot see on the page.
Named authors with a track record
Tips are advice, and search engines are cautious about advice in gambling. A named tipster with a profile, a history of their tips and a public record of how those tips performed is a far stronger signal than a byline that says "Staff". The record has to be honest, losses included. I wrote separately about auditing that record before anyone saw it.
A news sitemap that only holds news
Google News reads a separate news sitemap. It should hold only articles published in the last two days, up to 1,000 of them, with their publication times. Old articles belong in the ordinary sitemap. Leaving years of archive in the news sitemap does not help, and it makes the useful part harder to find.
What AI answers take, and what they leave
AI summaries at the top of results now answer a lot of simple questions directly. "What does evens mean?" is answered before anyone reaches a link. That is not something a site can fight, so the plan does not try to. It looks at what those answers cannot do well:
- Live prices. A summary cannot reliably show today's price with its age.
- Tools. A bet calculator, an odds converter or a calendar is something to use, not read.
- Accountable opinion. A named tipster's pick, with their reasoning and record.
- Records and data that only this site holds.
Those are the pages to invest in. Simple definitions are still worth having, for readers and as places to link from, but they are no longer where the clicks come from.
The short version
- Plan by search family, and write down which ones you will not chase.
- Give recurring topics one stable, current page.
- Index thin pages only once they clear a bar you have written down.
- Mark up only what is visible.
- Put real, accountable people on the advice.
- Build the things a summary cannot replace.
Further reading: Google's guidance on news sitemaps.