Websites/Affiliates/Meetings/2011-11-11

Attendees

Please add what worked and didn't work for team:

The challenges:

The initial delay in the Affiliates application is that the original timeline was created before WebDev was involved and it included zero days for development. The original timeline went from UX > Design > Front End > Zemoga support > Go Live. After WebDev was involved, a new timeline was created that included the back end development phase of the project.
There was also a three week days near the start of the project where we were waiting for the designs to be done. That was not in the original plan because Zemoga was going to do the designs and front end code and were available to do it.
There was also a delay during the initial phases of the project due to a product owner change.
WebDev only had one developer that had availability to work on the project.
Chris More took over WebifyMe and other projects during the initial phases of Affiliates and he became the sole technical project manager on the WebProd team leading to over-utilization of his time.
Documentation on all features and specs were not complete enough to move forward without asking additional questions.
Timelines were shifting and made it a challenge for all teams involved to also shift.

The Good
- The team was very open-minded to doing things different and making quick decisions. A lot of good insight on how to improve existing process came out of this.
- When push came to shove, we decided to delay the project past the quarter instead of release a very incomplete site. There was no flak, and everyone was immediately concerned with how to move forward, now how we had gotten to that point.
The Bad
- Desired functionality wasn't clear; the wireframes don't have enough detail to be the single resource for what the site is supposed to do.
- The admin interface for the site wasn't defined at all.
- The design did not match the wireframes. In several places items from the wireframes were changed or removed in a way that seriously affected the functionality of the site (the wireframes, for example, have a 'save' button on the 3rd step for generating a banner, but the designs had no save button). If this change was made intentionally, it was never communicated.
- Post-release, we have a lot of features that need UX work, l10n work, and other considerations before being developed; this isn't happening.

Challenges:

There were delays in getting the designs complete.
There were changes to the designs that were different from the original UX/mock ups.
The lack of site-wide navigation made it not extremely intuitive to move within the application.

The Good:

Appreciated dev + Engagement's hard work in testing the site together before QA even saw it
Also appreciate unit tests + early set up of critical traceback emailing lists

Security was a last minute review and most security bugs were pushed off to a 1.1 release to keep from relaying 1.0.

The application initially seemed slow and later was determined that it was just the daemons spinning up on the cluster. Expected behavior, but it caused a delay while we analyzed all of the code for improvements.

The good:
- The team was committed to making sure the site launched in multiple locales - a first for affiliate programs at Mozilla
The bad:
- The fact that locales needed to both localize the site copy in verbatim and the buttons wasn't clearly communicated - we still have some locales that have site copy but not buttons and vice versa
- buttons weren't available on dev for localizers to check that the right buttons had been uploaded