Our systems evolved over time to detect any anomaly in the rank tracking data and automatically deal with them before erroneous data could reach our clients.
How does this work?
The rank tracking system updates all the tracked keywords every day on mobile and desktop. There are four anomalies that the system detects and processes as designed, to avoid delivering erroneous data to you:
SERPs with over 100 results
The system detects if a percentage of keywords get SERPs with more than 100 results. In this case, Google seems to have rolled out a SERP update that affects our rank tracking process, and an intervention from the Engineering Team is required. The output:
A critical internal incident is automatically created, with the system calling the engineers on duty immediately (even if at night).
In the unlikely event of ranking data errors being delivered to clients, the Customer Success team publishes a public incident on status.seomonitor.com (always visible in the platform’s footer).
SERPs with less than 20 results
If a SERP contains less than 20 results (links), we retry 3 times, one of which with a 3rd party crawler. If the retries don’t work either, an intervention from the Engineering Team is required. The output:
A critical internal incident is automatically created, with the system calling the engineers on duty immediately (even if at night).
The newly-affected ranks don’t get recorded until the incident is closed.
In the unlikely event of ranking data errors being delivered to clients, the Customer Success team publishes a public incident on status.seomonitor.com (always visible in the platform’s footer).
Sudden rank drops from the top to 99+, on either desktop or mobile
When the rank suddenly drops from top 20 to 99+ on just one device. In other words, a significant variation in device discrepancies (one device on 99+, the other one in top 10) when the rank tracking on a market is done:
The newly-affected ranks don’t get recorded. If the discrepancy persists for more than 3 days, the 99+ rank gets updated (these 3 days help with possible "Google dance" situations and gives the engineering team enough time to fix a potentially broader #3 issue).
A critical internal incident is automatically created with the system calling the engineers on duty immediately (even if at night).
The Customer Success team publishes a public incident on status.seomonitor.com (always visible in the platform’s footer).
Sudden miss of a SERP feature, usually present
A significant number of keywords are reported as missing a specific SERP feature, usually present from one day to another. The output:
A critical maintenance task with a 48h SLA is automatically created.
Use-case
Google delivers an update to the design of the SERP, which affects our rank tracker’s ability to correctly recognize organic results and/or SERP features.
Other notes
There are situations when you could believe it's the ranking data that is not accurate when it actually is. This can be due to the time, the location, and the search parameters of your check.
For these situations, we created a Chrome extension that allows you to check your browser and what our crawlers extract, in real-time, side-by-side.