Moving from a Decentralized to a Centralized Tagging Implementation Model

[#1]

Background:

CarMax operates in a product team model, where each product team is responsible for a different area of the website. I provided the 1st party tagging requirements, and was responsible for all the administrative areas in Adobe, but each of those teams used their own developers to implement tags. Then we had another siloed 3rd party tag developer.

Operating in silos led to several problems:

1. Pipeline management

•Blocked pipeline delays critical updates

•Too many cooks in the kitchen

•Not best practice/goes against Adobe’s recommended approach

2. Risk

•Product dev’s don’t understand the implications of the changes they make, which puts at risk…

•All 3rd party media tags (power ~$xxM in media spend)

•1st party tags (power reporting, lead scoring models, etc.)

3. Tagging was deprioritized

•Product dev’s not all adequately trained on tagging

•Too many competing priorities

•No resources to support site-wide initiatives (like global tags)

4. Inconsistency

•Inconsistent use of Global Analytics

•Inconsistencies in implementation across teams

•Gaps in technology knowledge

•Led to low quality data

Pre-pitch:

Before making any kind of formal pitch, I took some important key steps:

1.Identified and built relationships with allies! (product team devs, 3rd party dev, technology leader, my manager, my VP, data science/acquisition team)

2.Organically socialized current issues with decentralization and promoted centralization with those allies; make sure they understood the vision

3.Worked with them to identify stakeholders/decision makers

4.Leveraged allies to organically socialize upward to their managers/stakeholders

5.Developed a vision, mission, purpose, and scope for a centralized team

6.Quantified centralization (provide ROI)

Gave stakeholders a pre-read of #5 and #6

Pitch:

1.Brought all the right stakeholders together

2.They already had context b/c of the organic socializing done by myself and allies

3.Stated that goal of meeting was just to get alignment; not yet asking for additional resources (but had slides prepped in Appendix if they did)

4.Official presentation/discussion (see appendix slides 21 – 40 for the actual deck I used)*

5.Tried to anticipate questions like how the team would operate, roles & responsibilities, # resources needed, etc. and had that ready in the Appendix

6.Pitch was successful! Gained unanimous buy-in from all stakeholders with green light to proceed forward and hire analysts and developers

Building the team:

1.Officially moved 3rd party dev onto our team/consolidated 1st and 3rd party tagging

2.Prioritized roles – Devs vs. Analysts vs. QE vs. Delivery Manager, etc.

3.Started hiring – prioritized skill set knowledge over tagging/Adobe experience (i.e. prioritized QA experience over tagging QA for QA role, because I could train in Adobe)

4.Lots of training and onboarding; lots of peer programming (hey let’s all do this together so we learn); retros and norm-setting as a team remains highly valuable

5.We were very scrappy during this time period and everyone on the team understood that it was a big transitional time and remained very upbeat and very flexible, wore a lot of hats, etc. Felt very startupy. Told people we hired that this would be the case – very transparent.

6.Grew from a team of 2 to a team of 12 (me, 6 devs, 3 analysts, 1 QE, 1 Delivery Manager)

Proving our Value:

1.I kept stats on our team’s efficiency, rate at which we moved cards, # of broken tags, etc. - As each new person came on, especially devs, I provided updates back to stakeholders – showed how all those metrics were improving

2.Interviewed product team devs and product managers to make sure we were fulfilling our promises of less tagging burden on them

3.Regular touch base with my technology counterpart (people manages devs)

4.SDK was a BIG accelerator to centralization – we are migrating over one microsite at a time to SDK, and once it’s migrated, DDS owns tagging from that point on

5.Monitor Core Web Vital trends pre and post SDK and share results – always huge lifts which increases demand for SDK/keeps our team always relevant

6.Take every opportunity to provide updates (Open Houses, etc.). Demand has never been as popular – stood up brand new tagging on 5 NEW properties in last 6 months

Key Takeaways:

1.Forge alliances and build relationships!

2.Do pre-pitch work so that by the time the pitch comes, it’s mostly a check mark of agreement from stakeholders

3.Put dollar signs against any ask – what is return on investment?

4.Build a team the right way with the right kind of people

5.Follow up after achieving goal to prove value – show that you’re accomplishing what you said you would accomplish!