Improve Radar
The short answer to how to improve your SEO is to score six axes honestly: technical health, on-page clarity, content coverage, internal links, external links and click-through. Then fix the lowest one first, for four to six weeks, and re-score before you touch anything else. Six-axis rebuilds fail because effort spreads thin and no single axis crosses the point where the engine notices. One axis lifted far enough raises the ceiling for the other five, which is why the order of work matters more than the size of the plan.
Below you get the six scores, one move per axis, and a three month target you can check without a dashboard tour.
No ranking promises. A scorecard and an order of work you can argue with.
Find the axis that caps everything else and move only that one for four to six weeks. In most audits it is technical health or content coverage, because both decide whether pages can rank at all. Score each axis from 1 to 5, fix the lowest, then re-score before touching the next.
A six-axis plan looks responsible. It is usually the reason a year passes with no visible change. Every axis gets a little work, no axis gets enough, and the site sits in a permanent state of being improved without ever crossing a threshold. Search engines reward depth on one dimension at a time far more than a thin spread across six.
The second reason is measurement. When six things change in the same month, you cannot tell which one worked. A ranking move gets attributed to the redesign, the new copy, the link buy or the speed ticket, and the next quarter repeats the guess. One axis at a time is slower on paper and faster in practice because every result is readable.
The third reason is team load. Technical fixes need a developer, content needs a writer, links need outreach, and click-through needs someone with design judgement. A plan that touches all six at once needs four people in the same fortnight, which is exactly when priorities get reassigned and the plan quietly slips to next quarter.
Use a 1 to 5 scale and write down the evidence next to each number. A score with no evidence is an opinion, and opinions do not survive a budget review.
Measure Crawl errors, index coverage, duplicate URLs, mobile field data, redirect chains. Pull one crawl and one coverage report on the same day so the numbers agree.
A low score looks like this: a fifth of the catalogue missing from the index, product filters creating thousands of duplicate addresses, and a mobile page that shifts while a buyer tries to tap the enquiry button.
Measure One primary query per URL, a title that states the offer, one H1, headings in reading order, and a description written for a human. Check ten pages by hand, not the whole site.
A low score looks like this: every page shares a template title with the company name first, and two pages compete for the same phrase without anyone noticing.
Measure The questions a buyer asks after the first one, and whether the page answers them without a follow-up email. Compare your page against the top three results for the same query.
A low score looks like this: specification tables with no compatibility notes, no failure modes, no lead times and no answer to the question that decides the purchase.
Measure Links per page, orphan pages, and click depth from the home page. A crawl plus a link report gives you all three in an hour.
A low score looks like this: new pages published with no inbound links, and the only path to a product page running through a menu that nobody crawls twice.
Measure Referring domains to the pages that need to rank, not to the home page. Quality beats count, and relevance beats both.
A low score looks like this: a handful of directory listings, no mention anywhere a buyer would actually read, and every link pointing at the home page instead of the pages that sell.
Measure Click-through rate at a stable position, not overall. A page at position six earning two percent is losing the click to the title below it.
A low score looks like this: correct rankings with titles that read like file names, and descriptions that repeat the H1 without giving a reason to choose you.
Pick one. Do it until the evidence shows, then re-score. The move column is deliberately narrow: a short list of changes beats a programme of work.
| Axis | The one move | Done signal |
|---|---|---|
| Technical health | Remove duplicate URL patterns and fix the redirect chains behind them | Indexed URLs fall to the intended set |
| On-page clarity | Rewrite titles and H1s on the ten pages closest to page one | Click-through rises at the same position |
| Content coverage | Add the three missing buyer questions to the five highest-intent pages | Impressions broaden into new query groups |
| Internal links | Link from the five pages with most traffic to the five pages that must rank | Orphan report clears, crawl depth drops |
| External links | Earn five relevant mentions for one commercial page, not the home page | Referring domains on that URL increase |
| Click-through | Rewrite descriptions as a reason to choose, not a summary | Click-through at stable position climbs |
One axis per month, one evidence file per month. Three months is long enough for a re-score to mean something and short enough that the team still remembers what changed.
Take the lowest axis from your scorecard and move it to a 3.5. Nothing else. If technical health is the floor, that means the crawl is clean and the index matches the page map.
Target: one report that a colleague can read without you explaining it.
Move the second-lowest axis to a 3.5 and start four to six pages that answer the questions your buyer asks next. Judgement, not volume: five good pages beat twenty thin ones.
Target: new impressions on query groups you did not rank for in month one.
Re-score all six axes with the same evidence you used in month one. Whatever is still lowest becomes the next quarter. Do not skip this step; it is the only place the strategy actually gets decided.
Target: a radar with one visibly longer spoke.
We score all six in a week, with evidence you keep, and hand you the order of work.
The scorecard only works if the underlying checks are honest. Two tools on a laptop can produce the technical and link scores; the content and click scores need a person reading pages like a buyer.
Run the checks in a fixed order so the numbers are comparable month to month. Our SEO Audit Checklist is the sequence we use: access first, then index coverage, then the page to query map, then links, then the click layer. Scoring out of order produces a radar that changes shape for reasons unrelated to your work.
If the weak axis turns out to be content coverage, the fix is editorial rather than technical, and it usually takes longer than people plan for. SEO Content Writing Services exist for that case: pages written against a query map, with the follow-up questions answered in the order a buyer asks them, and internal links placed while the draft is still open.
If all six moved a little, nobody learned anything. That is the pattern to avoid.
A laboratory equipment supplier scored 1 on external links and 2 on content coverage, and had spent a year rewriting product pages that already ranked on page two. The rewritten pages were better, and nothing moved, because the pages had almost no links and no answer to the question buyers asked before a quote: which model fits an existing bench.
We left the content alone for the first month. Instead we earned eleven relevant mentions for three specific product pages and linked them from the two application notes that already had traffic. The pages moved from page two to the lower half of page one within nine weeks, and the enquiries that followed named the application note, not the product page.
In month two we added the fit question to the top six pages. That was the content axis, done after the link axis had already changed the ceiling. Had both been done in the same fortnight, the team would have credited the copy and repeated the wrong move next quarter.
Every four to six weeks while you are actively working one axis. Re-scoring weekly creates noise because crawl and index data move slowly. Re-scoring quarterly is too late to catch a wrong move early.
Pick the one that blocks the other. Technical health blocks content because pages that cannot be indexed cannot rank. After that, content blocks links, because there is little point earning a mention for a page that answers nothing.
Not always. Many sites improve most by fixing titles and internal links on pages that already exist. New content is the right move when the query map has gaps a buyer would notice, not because a publishing calendar says so.
A full point on a five point scale over one quarter is a realistic target for technical health or click-through. Content coverage and external links usually move more slowly because they depend on writing and on other people.
Rarely, and not useful as a goal. A five on external links for a small exporter would take years. A 3.5 across the board with one axis at 4.5 is a stronger position than a thin four everywhere.
Score technical health and on-page clarity yourself with a free crawl tool and a spreadsheet. Fix duplicate URLs and templated titles first; both are free and both change what the index can do with your pages.
