You also need to DELETE the ANCHOR POINT keyframes when you are using transform to track an Object.
Because of what you want the text to do, this is a very advanced track.
Mocha is a planar tracker. You really need to track data from the right side plane and then the left side plane of the camera for the text to animate in the manner you illustrate.
Just tracking the front plane of the camera lens wont have it do what you want I think.
Tom Daigon
PrP / After Effects Editor
http://www.hdshotsandcuts.com
https://www.youtube.com/watch?v=DxPrG3WUyz8
(Best viewed at 1080P and full screen)
HP Z820 Dual 2687
64GB ram
Dulce DQg2 16TB raid