//Corresponding task from "add a link": {T277356}//
---
When we first deploy "add an image", we want to deploy it with a controlled experiment, somewhat similarly to "add a link". But the experiment will only happen amongst mobile users. Here are the details:
* With the advent of the images experiment, we want to end the link experiment.
** Ending the link experiment means making it so that all legacy accounts with Growth features and all future accounts with Growth features have this situation:
*** Replace the unstructured link task with the structured one. It will no longer be possible to do the unstructured one, though it can remain configurable in community configuration, in case a wiki wants to switch back.
*** If a wiki never had the unstructured link task because they had no applicable templates, then they should still get the structured one added.
*** Legacy users difficulty filter settings should not be changed, but new accounts should get the easy tasks by default ("copyedit" and "add a link").
** We could either choose a date to end the link experiment across all wikis, before we start the image experiment. Or we could end the link experiment as part of turning on the images experiment wiki by wiki. This can be up to engineers.
* For the image experiment, these are the groups that should be applied for account creations after the experiment starts:
** Group A** **Mobile A**: these users have "add an image" on as the only default taskcreated their account on mobile. Even though it's a mediumThey get "add an image" on as the only default task,. we're doing this so that we can do intensive learning on this task (40% of accounts).(40% of mobile account creations). This is the only group that will have access to "add an image" at all. Some of these users may subsequently visit their homepage on desktop. See this task about how we will prevent them from using "add an image" on desktop: {T295000}
** Group B** **Mobile B**: these users created their account on mobile. They have the default suggested edits experience without "add an image", in their difficulty filters at all. Their experience would include copyedit and "add a link" as their default tasks. so it would include copyedit and "add a link".(40% of mobile account creations)
** Group C: usual control group that doesn't have Growth features at all.**Mobile C**: these users created their account on mobile. They don't get the Growth features at all and are our usual control group. (20% of mobile account creations)
* For accounts created before the experiment starts,* **Desktop B**: these users created their account on desktop. They have the default suggested edits experience without "add an image" in their difficulty filters at all. Their experience would include copyedit and "add a link" as their default tasks. "add an image" should be added as a new medium task type(80% of desktop account creations)
** **Desktop C**: these users created their account on desktop. They don't get the Growth features at all and are our usual control group. (20% of desktop account creations)
* For accounts created before the experiment starts, but not selected by default"add an image" should //not// be added as a new task type. The users should just keep their existing difficulty settingsThis is actually different than "add a link", where we //did// add it for legacy accounts.