My Role
In this project, I was the lead Technical Program Manager. This means that I was responsible for estimating timelines, holding teams accountable to completing tasks, and navigating all the corporate nonsense that surrounds offshoring projects.
Context
A major office supply brand decided their costs were out of control to manage the creative marketing and production organization.
The hired McCann Worldgroup and Craft to come up with a cost effective offshore production hub.
We had to take over the studio and come up with a plan to save money.
What we did
- First, we inaugurated the studio we took over into the Craft family. We went to great lengths to make them feel like they were a part of the larger company.
- We deployed a standard set of the tools that we had in all our other studios. This would help them communicate and pass work between various locations in our network. The toolstack ensured that everyone was speaking the same technical language.
- Studios store a LOT of data. For every massive creative asset, there are hundreds of variants. We came up with a plan to create a bidirectional sync to send data between the India and US location. This meant that people felt like they were working off a single fileshare.
Difficult problems we encountered
Problem | Solution |
How do we transport petabytes of data from the US to India? | This one was tough, our estimates of how long it would take to sync all that data over the internet would blow out our timelines. We were planning to use a UDP protocol via Aspera to do the sync.
Instead, we exported all that data onto external drives, shipped them to India, loaded the data onto the data stores at our office, and flipped on the Aspera sync to sync over the changes. This saved us a significant amount of time. |
How do we train people to a US standard of quality? | Our offshore location had a great talent pool, but the talent on the US side was more agile. We had to train that agility into the offshore crew.
We scheduled our executives and champions to fly out to India non-stop to train and course correct the designers there. |
We erased terabytes of data - big whoopsie - how do we get it back? | Yeaaa. This was a tres commas situation before that was a thing. In our case, the mounting mechanism that we used on the Mac mini responsible for coordinating the sync dismounted the destination drives. The Mac thought the files were deleted and attempted to delete them at the source location.
We had to recover terabytes of data from tapes. BUT, we did observe that the delete syncs were ridiculously fast. Impressively fast, really. |
Different tools meant we were talking different languages. | We had to align the studio that we took over to use the technology stack we were using in Craft. There were attempts to integrate them but we made a hard decision to migrate folks to a new toolstack.
There was a lot of pain since many users were so ingrained with how they operated their processes. It was lots of training and onsite visits to course correct folks on how to use the toolstack. |
Lessons Learned
- Timelines are gonna slip in a complicated project like this - they always do. My job as TPM was to manage the expectations of when we’d reach various tollgates. Confidence was key here.
- You’ll never make everyone happy. When we took the existing tooling away from the people in the studio we took over, we nearly had a revolt. But it was the right decision to get aligned on tooling, otherwise you’ll always have an integration that needs managing and folks saying things like “That’s not what myyyyyyy system says on my end”
- Always have backups of everything. Never delete anything. Backups saved our skin several times throughout the project.