Do loop
Why Do SEO? Turn the Question Into a Four Move Loop
Why do SEO? Not to chase a ranking report, but to move one business number with work you can verify and stop when it stops paying. Teams fail this question in two ways: they ask it once at the start and never again, or they never define what finishing looks like. The fix is a loop with four moves, purpose then action then proof then exit, run every quarter rather than once. This page shows how to run each move, what evidence closes it, and the conditions that should make you stop spending.
Four moves. One number. An exit written before the work starts.
Four moves in sequence. Skipping one is how a budget quietly disappears.
Why Do SEO in One Sentence?
You do SEO to acquire demand you would otherwise rent, and you run it as a loop: choose the purpose, pick the smallest action that can move it, verify with data before repeating, and write the exit condition in advance so the work cannot drift into habit.
Before the loop
A quarterly loop beats an annual plan
Annual plans fail because they are written once, in a meeting, by people guessing. A quarterly loop assumes you will be wrong about something and builds a review into the process, so being wrong costs one quarter instead of one year. It also gives the work a rhythm that a team can actually keep.
The loop has four moves and no fifth. Purpose names the number. Action is the single change most likely to move it. Proof is the measure that decides whether to repeat. Exit is the condition that ends the work, agreed while everyone is still optimistic.
A working knowledge of the fundamentals helps here, but it is not the bottleneck. Read Whats Is SEO? if the vocabulary is still unfamiliar, then come back to the loop. The loop is what turns knowledge into a decision.
Baseline
Write three numbers on one page before spending anything: current non brand organic clicks per month, current enquiries from organic, and the value of one enquiry. Everything after this is compared against that page.
Test: a colleague can read the page and say whether this quarter worked.Go or no go
If the value of one enquiry is small and the purchase never repeats, the loop will not repay itself. Say so early. It is the cheapest decision in the whole process.
Test: value per enquiry times realistic monthly enquiries exceeds the cost of one cycle.The four moves
Each move is one question with one answer
Read these as a sequence. The output of one is the input to the next, which is why skipping a move does not speed the loop up; it just moves the failure later.
Which business number is this supposed to move?
Name one number, not a wish
If the answer is more traffic, the project is already unmanaged. Pick one: qualified enquiries, trial signups, revenue per session, or cost per acquisition. Write it with a baseline, a target and a date. Everything downstream depends on that sentence, because the same tactic can raise traffic and lower revenue at the same time. A page that ranks for the wrong intent does exactly that, and it looks like progress in every report except the one that matters.
Test: the number is countable in a tool you already have, and one person owns it.Which single action has the shortest path to that number?
Fix the earliest broken link in the chain
Crawl, index, rank, click, convert. A page that cannot be crawled needs a technical fix. A page that is crawled but never shown needs intent work. A page that is shown but not clicked needs a new title and a sharper promise. A page that is clicked but never converts needs proof: price, process, evidence, a human name. Choose the earliest broken link, because spending downstream of the break wastes the budget twice, once on the work and once on the delay.
Test: you can say which link in the chain is broken and show the screen that proves it.What evidence closes the move?
Measure the thing that moves first
Decide the measure before the work starts and set it at the level that actually moves first. Impressions before clicks, clicks before enquiries, enquiries before revenue. Measure revenue in week two and you will cancel a project that is working. Never measure revenue and you will keep a project that is not. The honest test is a before and after on the same query group, the same period length and the same device split, which is usually enough to separate a real move from seasonality.
Test: the metric can be pulled without asking a developer, and it was named before the work began.What makes this end?
Write the stop rule while you still like the project
Every loop needs an exit written at the start, when nobody is emotionally invested. It can be a date, a spend cap, a performance floor, or a change in the business. Written late, the exit is decided by sunk cost rather than evidence, and sunk cost always argues for one more quarter. Put the rule in the same document as the baseline, so the two are read together.
Test: someone outside the project could apply the rule and reach the same decision.Exit criteria
Stop conditions we write into every engagement
These are not marketing copy. They are the clauses that end a programme, and we would rather publish them than pretend they do not exist.
The site is not being reconsidered for anything, which points to structure or authority rather than content. Publishing more pages will not fix a site that is invisible at the impression level.
Action: re-diagnose before spending again.If organic costs more per lead than a well run ad account for two consecutive quarters, the channel is not earning its place in the budget. That is a fact, not a failure of effort.
Action: shift budget to the cheaper channel.A pivot, a merger, a discontinued line. The content asset is suddenly about the wrong thing, and patching it usually costs more than rebuilding the affected cluster.
Action: pause and re-scope, do not push through.The loop has no second move without a publishing queue. Paying for strategy that cannot be shipped is the most common waste we see in mid sized companies.
Action: pause the retainer, keep the plan.An algorithm update, a manual action or a change in how results are rendered can erase a position overnight. Re-diagnose the chain before assuming the same action still works.
Action: restart the loop at move one.Every client gets one condition of their own in the document. It is usually financial or seasonal, and it matters more than any of the five above because you will actually enforce it.
Action: written at kickoff, reviewed each cycle.One cycle
What twelve weeks look like in practice
A cycle is short on purpose. Long enough to see a signal, short enough that a wrong guess does not become a strategy.
- Weeks 1 to 2: diagnoseCrawl sample, index coverage, query inventory by intent, and the three competitor pages that beat you. Output: one broken link in the chain, with evidence.
- Weeks 3 to 7: buildOne action, executed properly. Rewrite the cluster, rebuild the template, or fix the crawl path. Output: shipped pages, not a slide deck.
- Weeks 8 to 10: measureCompare the same query group before and after. Impressions first, then clicks, then enquiries. Output: a direction, not a verdict.
- Weeks 11 to 12: decideRepeat the action, change it, or apply the exit rule. Output: a written decision with the reason attached.
Diagnosis
Use a checklist, not an opinion
Move two depends on an accurate picture of what is broken. That is why we start every cycle with a structured crawl rather than a conversation. The SEO Audit Checklist is the same one we work from internally, and it is ordered by the chain rather than by tool category.
If the audit finds nothing broken in the first three links, the honest conclusion is that the site needs authority or a sharper offer, not more pages. That is a useful answer, and it is cheaper to hear in week two than in month nine.
A written audit also makes the exit rule testable. Without a baseline, every conversation about stopping becomes an argument about feelings.
Failure modes
Four ways the loop breaks before the exit
Run one cycle with us
Twelve weeks, one number, one exit rule. If the loop does not move it, we say so and stop.
FAQ
Questions about running SEO as a loop
How is a loop different from a normal SEO retainer?
A retainer sells hours. A loop sells decisions: one purpose, one action, one measure, one exit per quarter. The deliverables look similar on paper, but a loop can end, and that changes how the work is chosen.
How many cycles before I should expect a result?
Technical fixes can show in one cycle. Content and authority work usually needs two to three before the trend is readable. If three cycles pass with flat impressions, the exit rule applies.
Can the loop run without a dedicated marketer?
Yes, if one person owns the number and a writer or developer can ship. The loop is designed for small teams precisely because it limits work in progress to one action.
What if the exit rule triggers early?
Then we stop and write down why. The plan, the crawl data and the content briefs stay with you, so restarting later costs less than starting cold.
Do you still promise rankings?
No. Nobody controls the index. We commit to shipping the action, measuring it honestly and applying the exit rule, which is the part a supplier can actually be held to.