What the data team you hired in 2022 looks like in 2026
The hires that made sense in the analytics-everywhere era make less sense now. The structural mistake is treating the team you have as the team you would build today.
The data team your firm built in 2022 was, in most cases, structured around a particular set of assumptions. Self-service analytics was the goal. Modern data stack was the architecture. Each function got its own analyst. There was a vague aspiration toward “ML someday”. The team scaled headcount in proportion to the dashboards being requested.
Four years on, very few of those assumptions are still load-bearing. The dashboards have not gone away, but they are no longer where the firm gets leverage. The tooling around analytics has compressed into a smaller number of more capable platforms. ML - actual production ML, not aspirational ML - is now where the differentiation happens. And the labour market for that work looks nothing like the labour market that staffed the 2022 team.
A short observation on each shift, written for founders and Heads of Data who suspect they have inherited a team shape that no longer fits the work.
I - Analyst roles need to be redrawn, not eliminated
The temptation, in the cost-conscious posture most firms are in, is to thin the analytics layer of the team. We have seen this go badly more often than it has gone well. The work has not disappeared; it has just stopped being the highest-leverage activity. The right move is rarely fewer analysts - it is analysts with a different brief.
The brief that works in 2026 is the analyst who functions as an embedded translator between a business unit and the data infrastructure. They write SQL fluently, they own a portion of the modelling layer, they own the metrics definitions that the business unit operates against. They are not “the dashboard person”. They are a quasi-product role with a data stack underneath.
The team designed in 2022 has too many of the former and too few of the latter. The cheapest correction is to give two of the strongest existing analysts the latter brief, with the authority to refuse work that does not fit it.
II - The ML team you actually need is smaller and more senior
The post-2022 ML hiring wave produced a lot of mid-level ML engineers in head-of-data orbits that did not, in the end, ship anything to production. The market correction that followed thinned that population - but it also produced a clearer view of what an ML team that ships actually looks like. It is smaller. It is more senior. It is built around one or two operators with a track record of shipping models into a production environment under real constraints, with the rest of the team a mix of data engineering and applied research.
For firms that hired generously into the ML function during the growth chapter and are now wondering why nothing has shipped, the question is almost never about effort. It is about whether the senior ML hire at the centre of the team has actually built and operated a production system before, or whether they have built one in a notebook environment.
III - The data engineering function is now the leverage point
The pattern we see most consistently in firms that are getting good outcomes from their data investment in 2026 is that the data engineering function is the strongest part of the team. The analytics layer rests on it. The ML layer depends on it. When data engineering is weak - unstable pipelines, undocumented schemas, opaque lineage - everything else compounds slowly.
This is also where the hardest hires sit. Strong senior data engineers in this market are well-paid, well-known, and rarely actively looking. They are also, in our experience, the role most likely to be under-briefed by a Head of Data whose own background is analytics or ML.
Founders and Heads of Data we work with in this sector are mostly doing one of two things in 2026. Either they are quietly re-shaping the team they inherited from the growth chapter, or they are running an external search for the senior data engineer or ML operator who will let them do that re-shaping. The firms that get the most leverage out of their data spend are not the ones that hired most aggressively in 2022. They are the ones that have been most willing to look at the team they have, write down what it would look like if they started today, and act on the gap.