The 21-point on-page SEO checklist, re-ordered by return

Zach (@zach_weup) sits down with a laptop, starts a timer at 1:00, and fires off twenty-one on-page SEO fixes before it hits zero. Captions pop up on the right, one per item. The framing is “$100k of on-page SEO advice in 60 seconds” — his number, not a price anyone paid — and it ends on “follow for more.”

The advice is good, and that is worth saying before anything else, because everything below is a criticism of the running order rather than the content. Do all twenty-one and your site would be in genuinely good shape. But a checklist with no order is a list of everything, and the actual problem — for a small business, or for us doing client work — is that you have a week, not a quarter.

The list, in the order he says it

1. Go to Google Search Console and fix any issues.
2. Add breadcrumbs, add canonical tags.
3. Fix your Core Web Vitals.
4. Fix orphan pages.
5. Add schema markup.
6. Fix your heading structure — H1, H2 tags.
7. Write original content, avoid duplicate content.
8. Get high-quality backlinks.
9. Use clean, descriptive URLs for each page.
10. Fix broken links and 404s.
11. Make sure JavaScript is crawlable.
12. Fix your keyword cannibalisation.
13. Make the site mobile friendly.
14. Add author bios and E-E-A-T signals.
15. Keep titles around 50–60 characters.
16. Write unique meta descriptions.
17. Merge thin or overlapping pages.
18. Optimise images and add alt text.
19. Match every page to the search intent.
20. Find high-volume, low-competition keywords.
21. Add internal links to important pages.

That is the whole video. Nothing on it is wrong.

Two things the format hides

The first is that it isn’t really an on-page checklist. Roughly two-thirds of it is technical SEO — crawling, indexing, site health — which at most companies is a different job done by a different person. On-page SEO is what is on the page: the words, the headings, the title, the match to what the searcher wanted, the links between your own pages. Core Web Vitals, orphan pages, canonical tags and JavaScript crawlability are all worth doing and none of them are on-page. Item eight, “get high-quality backlinks,” is off-page by definition — it is the single largest job in SEO, and it is sitting halfway down a list about page markup.

The second is the absence of any order. Twenty-one items delivered at the same volume in the same minute read as twenty-one equal items, and they are nowhere near equal. Fixing the heading structure on a page that answers the wrong question is tidying a room in the wrong house.

The same twenty-one, sorted by what pays first

Three tiers, sorted by one question: does this change what a page can rank for, does it change whether the page can be found at all, or does it change how much a person and a search engine trust it once they have found it?

Tier 1 — the ones that change rankings

Relevance work. This is where the first hours go, and it is the tier most likely to be skipped, because none of it shows up as a red flag in a crawl tool. A site can pass every automated check and still be answering questions nobody asked.

• Match every page to the search intent. The highest-return item on the list, and he says it nineteenth. If the person searching wants a comparison and your page is a sales pitch, no amount of schema markup rescues it.
• Fix keyword cannibalisation. Two or three of your own pages competing for the same query, splitting the signal between them so none of them wins it. Ordinary on service sites where each page was written by a different person in a different year.
• Merge thin or overlapping pages, and write original content. The same job from the other side: one page that fully answers something, instead of three that half-answer it.
• Headings, titles and meta descriptions. Three of his items doing one job — H1 and H2 structure, a title around 50–60 characters, a unique meta description per page — which is telling the reader and the search engine what the page is about before either commits to it. The meta description is not a ranking factor; it is the ad copy for your search result, and it changes whether the result gets clicked.
• Add internal links to important pages. The cheapest lever anyone owns and the most neglected one. Every backlink needs somebody else to agree; every internal link is yours to place this afternoon.
• Find high-volume, low-competition keywords. In this tier with a caveat: it is research, not a fix. It tells you which pages to build next; it does nothing to a page you already have. Do it after the five above, not instead of them.

Tier 2 — the foundation that makes any of it findable

None of this improves a page. All of it decides whether the page is crawled, indexed and understood — so tier one is wasted without it. It is also the tier most amenable to being done once and left alone.

• Work the issues in Google Search Console. The only place a search engine tells you directly what it thinks of your site, and it is free. Start here because it is information rather than guesswork.
• Fix broken links and 404s, and fix orphan pages. An orphan page is one nothing else on the site links to; if no page points at it, both a crawler and a visitor have to already know it exists.
• Canonical tags. One official address per page. This one bites harder than it reads, because a wrong canonical tag looks like nothing at all: the page renders perfectly while quietly telling Google the real version lives somewhere else. We did it to ourselves this month — two live domains serving one identical build, so every page announced the other domain as its true address, and Google believed it.
• Breadcrumbs, clean descriptive URLs, schema markup. Three ways of saying the same thing to a machine: where this page sits, what it is called, and what kind of thing it describes. Cheap to add, and they age well.
• Optimise images and add alt text. Two separate jobs wearing one bullet. Compression is a speed job; alt text is an accessibility one. If you only do one, do the alt text — a missing alt is a defect for a real person, not just for a crawler.

Tier 3 — trust, speed, and the one that isn’t on the page

Real, and slower to pay back. Do these once tiers one and two are done — unless one of them is visibly broken, in which case it jumps the queue. A site that is unusable on a phone is not a tier-three task.

• Core Web Vitals and mobile friendliness. With the caveat above: “could be faster” is tier three, “unusable” is an emergency. The gap between those two is worth measuring before you decide which one you have.
• Make sure JavaScript is crawlable. If your content only exists after a script runs, treat it as at risk of not being read — by some crawlers, and by the AI systems that increasingly summarise a page before anyone clicks through to it.
• Author bios and E-E-A-T signals. Experience, expertise, authoritativeness and trust — Google’s own shorthand for “is there a real, accountable person behind this.” It is not a score you can look up. It is a set of things that are either true about your site or not, which is also why it cannot be done in an afternoon.
• Get high-quality backlinks. Not on-page at all, and by some distance the biggest item on the list. It sits down here because it does not belong in either of the other tiers, not because it is low priority — it is a months-long job rather than a week-long one. We wrote separately about where a Canadian business can get the free ones, and what “free” is quietly doing in that sentence.

Why a good checklist travels badly

The Reel has twelve likes, which is not a verdict on the advice. Resource-style posts in the same genre — the ones handing over a spreadsheet, a template, one specific trick — routinely do many times better with thinner content. The difference is structural: a checklist is something you watch, and a resource is something you save and send to somebody. Sixty seconds of captions leaves you nothing to hold onto once the timer hits zero, which is roughly why this post exists.

Worth noticing if you make this kind of content yourself: checklist videos build the impression that you know everything, and give-something-away videos build an audience. They are different jobs, and posting the first while expecting the numbers of the second is the most common way to conclude that “the algorithm hates me.”

If you only have one week

In order:

1. Open Search Console and read what it already says. Tier two, but it is free information and it takes an hour.
2. List every page and write, next to each, the one question it should win. Any page you cannot finish that sentence for is your first real job.
3. Find the pages where you wrote the same question twice. Merge them, or make them genuinely different.
4. Fix the title, the H1 and the meta description on the pages that survived step three.
5. Add internal links pointing at the pages you just decided matter.

That is five of the twenty-one, and most of the return. The other sixteen keep.

What this is, and isn’t

This is a re-ordering of somebody else’s list, not a ranking guarantee, and no ordering survives contact with a specific site untouched. If yours is fast and well-structured and answering the wrong questions, your tier one looks nothing like that of a site nobody can find. The tiers also assume a site that already exists and gets some traffic; a brand-new site’s first job is having something worth indexing at all. We have no client results to report yet — the site says so on purpose — so nothing above is a claim that this order moved anyone’s rankings, ours or a client’s.

If working out which tier your site is actually in — and then doing the work — is something you would rather hand off, that is what our Content & SEO work covers. Get in touch and we will tell you honestly which of the twenty-one is worth your time.