What Changes the First Time You Lead a Data Team

A couple of years ago I picked up The Chief Data Officer’s Playbook by Caroline Carruthers and Peter Jackson. I read it for what it had to say about leadership and making the right impact, and it connected in ways I didn’t expect. The book is written for someone walking into a single organization to build and lead a data function from the inside. I was doing fractional consulting, moving between clients, building foundations I’d eventually hand off. The contexts were different. The disciplines were the same.

Reading it felt therapeutic, honestly. I could see past projects and client challenges mapped to frameworks that gave them names. Moments where I’d gotten the sequence right without knowing why. Moments where I hadn’t and could now see what I’d missed. The book didn’t tell me what to do next. It helped me see the shape of the work I was already doing.

What I keep coming back to is how much of Carruthers and Jackson’s thinking applies outside the CDO role. Two areas in particular: how you approach a consulting engagement from day one, and how you navigate the transition from individual contributor work to something more strategic.

The translation layer

Carruthers and Jackson call communications the first of their six secret ingredients for data leadership. Not technical communication. Translation. The ability to take the same problem and describe it differently for the board, for IT, for operations, for the data team. They’re specific about this: storytelling and passion are the must-have abilities.

I felt that on a call last year where I was working with a client’s CDO on one track, their CMO on another, and then sat down with a board member from the same company. Those conversations give me energy because I understand their problems, and then the question is how to translate that to the technical side. The same data quality issue is a compliance risk at the board level, a workflow problem for marketing, and a schema decision for the data team. The CDO’s job, and the consultant’s job, is to hold all three translations at once.

Francesco Mucio and I were talking about a consumer goods brand a while back. Eighteen months of Shopify data, twenty-one million in revenue, half a million orders, AOV around forty-two dollars. An 87/13 split between one-time and repeat buyers, which is typical for consumer goods at that stage. I told him that moving even five percent of the one-time cohort into a repeat purchase is roughly a million dollars a year at their economics. More if the repeat cohort has better LTV than the first-purchase cohort, which it usually does.

That conversation was entirely about the revenue number and the math behind it. Not dbt models, not semantic layers. The Playbook would call that the translation from data to knowledge to wisdom. At some point you stop treating the technical work as the deliverable and start treating the decision it enables as the deliverable. When the decision doesn’t matter, the technical work is a hobby.

The first hundred days, applied to consulting

The Playbook has a detailed framework for a CDO’s first hundred days. Listen first. Ask stakeholders what keeps them up at night, not what data problems they have. Run a maturity assessment to anchor the conversation. Identify quick wins but don’t promise them on day one.

I think about that framework every time I start a new engagement. The instinct early in my consulting work was to start with the technology. Map the stack, identify the gaps, propose a migration. The proposals were technically sound and went nowhere, because the people who would have to fund them didn’t yet understand why any of it mattered.

A client hired a CDO partway through one of my longer engagements. I remember telling a colleague that we were solid for about 120 days because it would take the new CDO that long to get acclimated. That number came straight from the Playbook’s framework, and it was almost exactly right. The CDO spent the first few months listening, building relationships, and figuring out what the org actually needed before making any structural changes. I watched that process from the outside and realized I should have been doing a version of it at the start of every engagement.

The first move I make now is to find the decision the executive is losing sleep over. Not “what data problems do you have,” which gets a list of complaints. Not “what dashboard do you want,” which gets a wish list. Closer to “what is the call you are trying to make this quarter that you don’t have good information for.” That question comes directly from the Playbook’s first hundred days thinking, and it changes the shape of the engagement from the start.

Credibility without authority

There’s a line in the Playbook that has stuck with me: credibility is better than authority, and authority should be delivered through credibility. Carruthers and Jackson are talking about the CDO navigating an organization where everyone technically owns some piece of the data but nobody owns it all. The CDO has to earn the right to coordinate across those boundaries.

In fractional consulting that dynamic is even sharper. You have no hierarchy to fall back on. Last September I was on a weekly call with Rick at a client. Rick was about to escalate a custom JSON tagging build to the backend engineer for a segmentation problem. The right impulse from inside his frame. From where I sat, it was a new work stream for something we could probably solve with features already in the tool we were using.

I pushed back. Not hard. Before we pull a backend engineer in, let’s ask whether the existing topic model can do it. Ten minutes of working through custom affinities inside the platform, and we had a working approach that didn’t need a separate build.

I didn’t have the authority to tell Rick no. He’s a peer. What I had was enough credibility from the work we’d done together to slow the conversation down for ten minutes instead of letting the decision harden. That’s the Playbook’s credibility principle in practice.

The harder version of this shows up during cost-cutting cycles. I’ve watched a CDO I work with closely have to justify my engagement to a CEO focused on cutting costs. That conversation is about credibility at its most compressed. If the work hasn’t already earned its keep in ways the CEO can see, no amount of explaining will save it.

Building capability in a team you’ll leave

The Playbook spends real time on building the CDO’s team, and one of the things Carruthers and Jackson are honest about is how much harder it is than the advice makes it sound. Working with willing volunteers and enthusiastic amateurs, they write, will always outperform press-ganged experts.

A year ago I was running a recurring audience meeting that was taking more time than I could justify. I announced I’d shift it to office hours and have Adrian, a contractor, facilitate. The room agreed.

It didn’t stick. Adrian could run the logistics but wasn’t close enough to the architectural questions, and stakeholders kept addressing their questions to me. I quietly stayed on. A few weeks later I reset Adrian’s scope in a 1:1, because I’d handed him a role he wasn’t set up to hold.

The Playbook’s framing helped me see what I’d gotten wrong. I’d delegated the meeting without delegating the knowledge. The CDO version of this is building engagement champions across the business, small armies who spread the message rather than one person trying to hold it all. I hadn’t built the champion. I’d just reassigned the calendar invite.

Vertical strikes over horizontal layers

Brandon Nelson said something to me on a late call last year that I’ve been thinking about since. His framing: the engineering mindset gets stuck on building the right thing first, when the move is to prototype something with loose edges, have the conversation, and let the conversation tell you what to build. The idea isn’t new. But Brandon was saying it about a project I was on, and it took a few days to work out why it landed differently this time.

The Playbook has a framework for this. Carruthers and Jackson describe what they call the vertical strike strategy. Instead of building the data foundation horizontally, layer by layer, you pick a contained use case and strike vertically through all four layers: raw data, cleaned and contextualized data, knowledge, and the wisdom-level decision at the top. You demonstrate the full value chain in a small area, then run adjacent strikes that build on each other.

Brandon was describing the same thing in different language. Don’t build the whole foundation before you show the value. Pick one slice, take it all the way to a decision, and let that decision earn the credibility to build the next one. The CDO who tries to build the entire data foundation before showing any value gets defunded. The consultant who does the same thing gets replaced.

I think the Playbook connected with me because it gave names to things I’d been learning by feel. The frameworks aren’t prescriptive. They’re diagnostic. They help you see what you’re already doing, where you’re doing it well, and where the gaps are. For someone navigating the transition from building things to guiding how things get built, that’s the kind of book you keep coming back to.