Why do my survey points all have the same XY?

[AdamDaily]

I had a field crew out collecting survey data in an area with low connectivity to our network. There is a mandatory geopoint field that must be collected to submit a survey. Our field crews have been properly trained on how to capture locations using Phone GPS or manually setting and locking in a location. Due to slow network connections, they collected multiple surveys and saved them to send later. Once back in the office, surveys were sent from the outbox. So far so good.

When I went to review the data however, all of the points, that should be scattered over an entire building, ended up in the exact same location with identical latitude and longitude coordinates. On the map lit appears to be only 1 point. When I select this location, you can see the points are stacked and there are actually 11 features here. Anyone else seen this or know why it is happening? We have thousands of features to collect and this could potentially be a big problem. 

[RobertBorchert]

You were using a phone to collect GPS points inside of a building?

Chances are your phones could not get an accurate GPS point so they simply used the generic location.

[AdamDaily]

Our field staff are aware that GPS will not work in some locations. We have side loaded basemaps for them to reference in these areas and have trained them on how to manually place a geopoint and lock in a location when GPS is not working. Its also hard to think that the GPS never once updated or moved positions and locked in on 1 XY the entire time they were collecting. 

I feel it may have something to do with saving completed surveys to send later when network connections are slow or unavailable. We have disconnected editing / sync turned on and our field crew did say they saved each survey to the outbox and then sent everything at once later on.

I don't know why or how survey would reset geopoint XY on all of the surveys, but that could explain the same XY showing up on all of them. I've seen some strange behavior in survey before where previously collected surveys when reopened lose a lot of the data/attributes that had been collected.