As a very active editor of the Google Maps database, I have had a suspicion that not all my approved edits get counted and this results in earning fewer points. I have had this suspicion for 2-3 years without caring much about it. And I did not want to spend the time needed to test and document this properly. You know, points are not the main driver for our contributions 
Very recently I finally reached 600K points. This motivated me to initiate a final editing effort to reach the target and to celebrate.
Here is a graph showing my points total in the days before and after I reached 600K points earlier this month.
To document potential bugs in the system, there are a few issues to understand. These include:
When you look at this graph there are two distinct jumps: Between January 14th and 15th there is a jump of 1895 points and between January 22nd and 23rd the points jump by 3470 points. This is not a reflection of me making contributions in huge bursts, but because total points are only updated once a week or so. In this case, 9 days went by between the major recounts.
As you can see my total points did not move at all from January 19th till January 22rd. It would be a lot easier to check our points if they were counted and updated daily. So this post is based on the assumption that points and contributions made between the jumps are always included in the following recount.
Counting email notifications
To document how many edits were made I decided to rely on the email notifications we receive after submitting edits. We donât always get such notifications, but it is the best I can do. Counting contributions in the contribution list is less accurate for many reasons.
In Gmail, I have a simple filter to move all such email notifications into a folder called âmaps thank youâ. In this folder, I made some advanced searches as shown below.
In this example, I searched for email notifications made on January 19th. Notice the search string underlined with red. In this search string, both dates need to be edited manually to ensure only emails from a specific date are shown. Then you need to hit â>â repeatedly until you reach the last page of search results. In this case, 100 emails are shown per page, and on the last page, the total number is shown. See the red circle: I made 814 successful edits on January 19th.
This number is probably slightly on the high side, as edits submitted in the prior weeks and approved on January 19th are also included. A few road edits have incorrectly been counted as edits.
Also, when more than one edit is done to the same pin, these are counted as âoneâ in Gmail, as shown below because the counter is for lines and not for the number of emails.
When making edits I switched between using English and Danish, so it was necessary to count the email notifications using two different search terms: âPublishedâ and âOffentliggjortâ which is Danish for published.
Here are the results:
When looking at the first graph it is reasonable to expect that the contributions made on January 15th through 22nd will be included in the points jump shown on January 23rd. These are the numbers marked in red. The sum of the edits marked in red is 1347 edits. As one edit equals 5 points, I would expect the points to increase on January 23rd by at least 6735 points (1347 edits * 5 points/edit).
As already pointed out the points jump on January 23 is only 3470 points. So I tend to conclude that 3265 points are missing. 3265 out of 6736 is 48.5% that is missing.
What can explain this?
- Are subsequent edits to a pin not counted?
My answer: This can not explain the 48.5% discrepancy.
- Do the points recounts not include the dates immediately before the recount?
My answer: Highly unlikely in my opinion.
- Are there further unknown delays after recounting contributions to publishing the points?
My answer: Highly unlikely in my opinion. I will check after the next recount in a week or so.
- Are the edits not counted correctly?
My answer: Probably!
- Could it be that the fewer points are caused by some photos and reviews being removed by the spam filter or because the businesses closed?
My answer: Yes this could be. Not very likely though.
- During this period the spam filter wrongfully blocked my ability to make edits on Google Maps. But this was reversed within a few hours of my appeal. Could this be the cause?
My answer: Probably not because none of my edits disappeared as far as I could see. This issue has persisted for 2-3 years as mentioned above.
To test items 4 and 5 I decided to also analyze the contribution stats for edits over the same period. Here are the results:
The total for edits marked in red is 825 edits from January 15th to January 23rd. This is the period I would expect to see reflected in the following recount update between January 22nd and 23rd.
Based on email notifications from the same period, the number of edits should be 1347. So what happened to the missing 522 edits (1347 - 825)? This means that 39 percent of my edits were not counted.
It is difficult to explain why 39% is not 48.5% but it could be related to other contributions being made or removed in the same period.
But still losing 39% of your contributions due to a counting bug is very discouraging. Especially, when you consider that these are only the approved edits. Quite a few of our edits do not get approved.
Conclusion
To keep volunteers happy, engaged, and motivated to improve Google Maps by making accurate edits, I suggest this bug be investigated and fixed.
Maybe I made mistakes in these calculations. I would be happy to learn where I failed and Iâm ready to make the needed changes to this post.
The results are summarized in this image:
Cheers
Morten
