Know where you rank, and which page is doing it

A rank tracker answers one question: where does this site appear in Google for the searches that bring in business? The hard part is answering it consistently. Google results change by location, by device and from one day to the next, so a position you check by hand at your desk tells you very little about what a customer across town sees. This article explains how the RankThat rank tracker gets its numbers, what each figure on the screen means, and how to turn it into a routine that clients understand.

If you are new to the idea, the glossary entries on rank tracking and average position are a good two-minute primer.

Setting up tracking for a site

Open a project and go to SEO › Rank Tracker. With nothing tracked yet, click Add keywords to track. You can also send keywords here from Keyword Research, Keyword Magic or a saved keyword list, which is how most agencies do it after the first research session.

Then open Tracking settings. Four choices matter.

1. How often. Weekly (the default), three times a week (Monday, Wednesday, Friday) or daily. Weekly suits most local businesses. Daily earns its cost during a site migration or right after a big change.
2. Devices. Desktop, mobile or both. If the client’s customers search on phones, and for most local trades they do, track mobile.
3. Location. Search for a country, state or city. A keyword can have its own location if one service area differs from the rest. Changing the location makes the next check a new baseline, because results from a different city aren’t comparable.
4. Competitors. Add up to 10 competitor domains. Their positions come from the same search results, so they cost nothing extra.

Scheduled checks run every morning, and Check now runs a check on demand.

Choosing what to track

Track the searches that map to money: each service, each service with the city, and the handful of questions customers ask before they buy. Leave out branded searches unless the brand is new, since a business almost always ranks for its own name. Tag keywords by service as you add them. Months later, being able to filter “water heaters” against “drains” is what makes the data useful in a client meeting.

What you see on screen

The top of the page has a toolbar: the desktop and mobile toggle, date presets (7, 30 or 90 days) or any two dates you choose, a tag filter, and buttons for Clean up for reports, Tracking settings and CSV.

Below that, a row of summary tiles:

A “Visibility vs competitors” strip follows, and then three views of the data: Keywords, Landing pages and URL switching. At the bottom, a chart shows visibility over the last 12 weeks.

How visibility, traffic and share of voice are worked out

All three use the same idea: a result at position 1 gets far more clicks than one at position 9. RankThat uses a standard click-through curve (about 32% of clicks at position 1, about 3% at position 10, about 1% on page two).

• Visibility is the expected click rate at each keyword’s position divided by the rate at position 1, averaged across your keywords.
• Est. traffic is each keyword’s monthly search volume times the click rate at its position.
• Share of voice is the clicks you would earn at your positions divided by the clicks you would earn ranking first everywhere, weighted by volume. Competitors are scored the same way, so the comparison is fair.

These are models, not measurements. For real clicks, look at Search Console data. The tracker tells you about position; Search Console tells you about traffic.

Checks that keep the numbers honest

Rank data is noisy, and a tracker that reports every wobble generates false alarms. A few things happen behind the scenes:

• Big moves are checked twice. If a keyword disappears or moves 10 or more places, it is checked again and the better result is kept. Google sometimes serves an odd page of results, and one bad sample shouldn’t become a client email.
• Short result pages count as unknown, not as a drop. If Google returns a shorter list than usual and your page isn’t in it, the position is saved as unknown.
• Positions are organic positions. Ads and the map pack don’t push your number down. The tracker reads the top 100 results.

Clicking into a keyword

Click any keyword to open its history: position over 30, 90 or 180 days, the best position it reached, the page that ranks, and the top results around you. This is where most “why did we drop?” questions get answered. Often the page didn’t fall; Google swapped it for a different page on the same site.

That swap has its own view. URL switching lists keywords where more than one of your URLs ranked during the period. It is a common sign of keyword cannibalization: two pages competing, neither winning. The fix is usually to make one page clearly the best match and link the other to it.

Alerts for drops and wins

You don’t need to open the tracker daily to catch problems. By default, RankThat raises an alert when a keyword with at least 30 monthly searches drops 10 or more places or falls out of the top 10, and a separate “win” alert when a keyword reaches the top 3. Both thresholds can be changed per project. Alerts arrive in a daily digest by email or Slack, and they appear on the Alerts page across all your sites.

A worked example: Brightside Dental

Brightside Dental has three locations in one metro area. The agency tracks 60 keywords on mobile, weekly, with each location’s keywords set to that suburb (“dentist in Round Rock”, “emergency dentist Cedar Park”). They add the two practices that show up most often as competitors.

In week six, an alert says “teeth whitening cost” fell from position 5 to position 18. The keyword history shows the ranking page changed from the whitening service page to a two-year-old blog post about whitening at home. URL switching lists the same keyword. The agency adds a clear link from the blog post to the service page and rewrites the service page’s title from Opportunities. Two weeks later the service page is back at position 6. Because the title change was made through RankThat, it is recorded, and Opportunities measures the clicks before and after.

Cleaning up keywords before they reach a report

Every tracked list collects junk over time: typos imported from Search Console, competitor brand names, searches for other cities. Clean up for reports labels these as typos, duplicates, competitor names, other locations, not relevant or junk. They stay tracked (you might want to watch a competitor brand term), but they are hidden from client reports. The monthly report keyword table can then show only keywords that rank, or the full table, without someone deleting rows by hand.

Mobile and desktop positions often differ by a few places, and for local searches the gap can be larger. If you track both, report the one that matches how the client’s customers search, and mention the other only when they move in opposite directions.

A monthly routine for agencies

1. Week 1: read the alerts digest. Open any keyword that dropped 10 or more places and check URL switching.
2. Week 2: compare share of voice against competitors for the last 30 days. If a competitor gained, look at their pages in Competitor Analysis.
3. Week 3: add new keywords from fresh research. Tag them by service so you can filter later.
4. Week 4: run Clean up for reports, then let the monthly report pull the table.

What the rank tracker doesn’t do

• It doesn’t track the map pack. Map results change every few blocks, so they need a grid of searches. That is what the Local Rank Grid is for.
• It doesn’t go past position 100.
• It won’t match every other tool exactly. Different trackers sample at different times and places. Watch the trend, not a one-place difference.
• It isn’t a traffic report. Estimated traffic is a model. Real clicks come from Search Console.

Tip: track fewer keywords well. Forty keywords that map to real services, tagged and cleaned, tell a client more than 400 keywords nobody reads.