The short version
- On-page is everything you control. Off-page is everything you influence. That difference in control drives every practical decision here.
- On-page is cheaper, faster, lower risk and has a ceiling. Off-page is expensive, slow, riskier and has almost none.
- Fix clear on-page blockers before scaling outreach. Content, technical maintenance and relationship-building can then run together.
- Ranking position alone cannot diagnose the constraint. Check indexing, intent, content, competition and links before choosing the work.
The division between on-page and off-page is the oldest framing in SEO and it survives because it maps onto something real: the line between what you can change this afternoon and what you have to earn from other people over months. Almost every practical prioritisation question reduces to working out which side of that line your actual problem sits on.
Most teams get this wrong in the same direction. They assume they need links, because links are the glamorous part of SEO and the part agencies sell, while their own pages have duplicate titles, no structured data and a heading hierarchy that makes no sense. That is spending on amplification before the signal is worth amplifying.
The short answer
Fit for your workflow
Start with demonstrable blockers, not a universal on-page-first rule.
A bounded on-page repair project can finish; content and technical maintenance continue as the site changes. Compare the expected benefit of a specific repair with a specific outreach project. Neither category has a universal cost or guaranteed ranking return.
- You do not rank at all for your terms
- Check indexing and query relevance first; competition and links may also matter.
- You rank on page 2 consistently
- Compare intent, content, technical signals and link evidence. Page two is not proof of a link deficit.
- You rank well but traffic is falling
- Check demand, positions, click-through and result layout, including AI Overviews.
- New site, no history
- Build useful, accessible pages while earning relevant mentions. No fixed waiting period applies.
What belongs to each
Look at the control column. Everything in the top half you can change today without asking anyone. Everything in the bottom half requires persuading a third party to do something that benefits you. That asymmetry is why on-page work has a dramatically better cost-to-outcome ratio right up until the point where it is finished.
A diagnostic that takes ten minutes
- Open Search Console and find queries where you get impressions but few clicks at low positions.
- For each, check: does a page on your site clearly and specifically address that query? If no page is genuinely about it, that is on-page.
- If a good page exists at a low position, compare intent, content, freshness, indexing and relevant links. An established competitor is evidence to investigate, not an off-page diagnosis.
- Crawl your site with the free tier of Screaming Frog. Duplicate titles, missing descriptions, redirect chains and orphan pages are all on-page and all fixable this week.
- Prioritise the largest evidenced constraint; relevant outreach need not wait for every minor on-page issue.
Most sites that run this find several quarters of on-page work they had not noticed, because on-page problems are invisible from the front end. A page can look perfect to a visitor and still have no structured data, a generic title, no internal links pointing at it and a heading hierarchy that says nothing about its subject.
The ceiling on each
Fixing a template has diminishing returns, but content and user needs keep changing. Strong pages can still face tough competition. Investigate the remaining gap rather than assuming that links are the only lever.
Off-page has almost no ceiling and a much worse cost curve. Each additional link or mention costs roughly what the last one did and delivers diminishing marginal benefit. That is why the sequence matters so much: exhaust the cheap finite work before starting the expensive infinite work. See backlinks vs content quality.
Which half we are useful for
Our tool works in your codebase, so it addresses on-page and technical work: titles, markup, structured data, internal linking, heading structure. It does nothing for off-page. We have no link index and no way to earn you a mention. If your diagnostic above points off-page, this is not the tool for that problem.