Your first audit should not start with a ranking checker
Author : Caleb Hawkins | Published On : 18 Oct 2026
The tab is still open from last Tuesday. You paid for a month of rank tracking, watched a green arrow appear next to three keywords, felt briefly excellent, and then closed the laptop without changing a single page on your site. That is the whole problem in miniature.
A google ranking checker tells you where you stand. It does not tell you what to do, and it certainly does not tell you whether the position it reports is real. Most of the money wasted on rank tracking comes from treating the number as a verdict rather than as a clue. What follows is the order I would do things in if I were starting again, plus the specific ways each step drains a budget.
Step 1: Decide what a ranking is even supposed to mean for you
Before you type a keyword into anything, write down the job that keyword is meant to do.
There are roughly three kinds of queries you will track, and they behave differently:
- Money queries where someone is close to buying or booking. Few searches, high value, brutal competition.
- Research queries where someone is comparing approaches. Higher volume, weaker intent, easier to rank for.
- Brand queries where someone already knows your name. These rank first almost automatically and tell you nothing.
Tracking all three in one dashboard is how people end up with a meaningless average position of 14.3 and no idea what it means. Split them. If a research query sits at position 6 and a money query sits at position 40, you have not got "mixed results". You have one page that works and one that needs rebuilding.
Watch for: tools that default to national-level tracking when your customers are local. If you serve a single city, a nationwide position is close to fiction. Set the location explicitly, and set it once, because changing it mid-month resets your comparison baseline.
Step 2: Check the checker before you trust the chart
Say you run a small equipment repair shop and you sign up for a rank tracker on a trial. It shows your main service keyword at position 4. You screenshot it, send it to your business partner, and start planning the extra work that "top 5" implies. Two weeks later a different tool says position 11. Nothing on your site changed. What happened? If you want to understand why tools disagree, what to know about the keyword position tracking is a useful resource to read.
Several things, all boring, all expensive if you ignore them.
Rank checkers sample the SERP from a limited set of locations. They may or may not personalise. They may count ads, map packs, or featured snippets differently. Some count the local pack as three results, some as one. Some report the highest position your domain appears at anywhere on the page, including a subpage you forgot existed. Others report only the exact URL you entered.
So when you compare two tools, you are often comparing two definitions, not two measurements.
What to do: pick one tool, note its settings, and never compare its numbers to a screenshot from a different tool. If you must switch, run both for a full month and record the offset between them. A consistent gap of six positions is annoying but usable. A random gap is not.
Watch for: free checkers that query once, from one server, with no location setting. They are fine for a rough sanity check on a brand new page. They are not a basis for keyword position tracking across a client portfolio, and they are a terrible basis for a report you send to someone who pays you.
Step 3: Build a baseline you can actually defend
Here is where most first audits quietly fall apart. People track twenty keywords, look at them weekly, and never write anything down in a form that survives a redesign, a migration, or a change of tool.
Build a simple sheet instead. One row per keyword. Columns for the query, the target URL, the intent group, the current position, and the date you recorded it. Add a column for the page that actually ranks, which is frequently not the page you wanted.
Then leave it alone for two weeks.
That pause matters. Ranking data is noisy. A page can swing five positions on a Tuesday because a competitor published something, because a SERP feature appeared, or because the tool's sample rotated. If you react to every swing, you will spend your budget rewriting titles that were never the problem.
Watch for: tracking keywords with no page behind them. If no URL on your site targets a query, its position is whatever Google decided to show, often a blog post from years ago. Fixing that is a content decision, not a tracking decision.
Step 4: Read the SERP, not just the number
Position 7 in one SERP and position 7 in another are different jobs.
Open the query in a private window with the location set correctly and look at what surrounds your result. If the top four slots are all comparison articles and you are selling a service, you are not losing on rank. You are losing on format. No amount of position tracking will fix a mismatch between what you published and what the page is rewarding.
Three things to record every time you check:
- Which result types occupy the top of the page (snippets, video, local pack, shopping units).
- Whether your ranking URL matches the intent of the query.
- Whether the pages above you are older, thinner, or simply better structured than yours.
That third one is the useful part. If a competitor with a worse page outranks you, look at their internal links and their page depth. If a competitor with a better page outranks you, look at your own outline and examples.
Watch for: chasing a position you cannot hold. If a query is dominated by large marketplaces and you are a five-person operation, position 9 may be a genuinely good outcome, and the money is better spent on a narrower query where you can reach the top three.
Step 5: Turn the chart into work
Rank data becomes valuable the moment it changes what you do that week.
A workable routine looks like this. Every second Monday, export positions. Flag only keywords that moved more than three places in either direction. For each flagged keyword, open the page and check one thing: does the page still answer the query in its first screen? If yes, leave it. If no, rewrite the opening and the subheadings. Log the date you changed it and check again in three weeks.
That is it. One change per flagged keyword, one recheck, no panic.
Watch for: treating SEO reporting as the deliverable. A monthly report full of arrows is not progress. If the report does not say which pages changed and what happened afterwards, it is decoration.
Where the money actually leaks
These are the recurring mistakes I would warn anyone about before they spend a second month on tracking.
Paying per keyword when you only care about five. Seats and keyword limits scale fast. Trim the list to queries that map to revenue or to a clear next step, and archive the rest.
Tracking daily. Positions do not move meaningfully every day, but your anxiety does. Weekly is plenty for most sites. Monthly is enough for a small local business.
Ignoring the target URL column. Half of all "we dropped" panics are actually "a different page now ranks for this query". That is not a loss. That is a signal to consolidate.
Reporting rank without context. A number in isolation invites arguments. A number next to the page, the intent, and the date invites decisions.
Buying a tool before you have a list. The tool is cheap compared to the hours you will spend staring at it. Write the keyword list first, with intent groups, then buy the smallest plan that covers it.
Forgetting that rankings are a lagging indicator. Traffic, conversions, and assisted sales are the outcomes. Rank is a way of explaining them, not a goal in itself.
Letting one person own the data with no notes. When they leave, the history leaves with them. Keep the sheet shared and the notes plain.
What to do when the numbers look wrong
They will look wrong. A keyword will jump from 30 to 3 overnight, or vanish entirely, and you will assume a penalty.
Check four things before you change anything. First, confirm the URL still returns a normal page and is not blocked. Second, confirm the location and device settings have not shifted. Third, check whether the SERP itself changed shape. Fourth, check whether the query is seasonal, because a query about winter tyres behaves differently in July than in November.
If all four come back clean, wait two more checks. Most anomalies resolve. The ones that do not tend to be real, and by then you will have saved yourself three unnecessary rewrites and one very awkward conversation with whoever approves the budget.
The habit worth building is simple. Check on a schedule, record what you saw, change one thing, and look again. The tool is a thermometer. You are still the one who decides whether the room needs a window open.
