By Jane Steen, Head of Engagement at ETL.
Last week, we put a tablet on our stand at the Data Decoded conference in Manchester and asked data practitioners a question: what actually derailed your last data project? 102 people answered, anonymously.
We found that only 6% blamed the tooling.
It was the least-chosen of the five answers on offer. The most common was waiting on another team or supplier, at 29%, with requirements that kept moving just behind at 28%. A further 21% said that the data had turned out to be far worse than anyone had admitted. Only 16% said that nothing had gone wrong.
Put the first two together and 57% of these projects were sunk by coordination issues. The technology – the part with a budget line and a procurement process – accounted for just 6%.
This graph summarises the results:

What the question didn’t ask
One caveat, because it affects how the chart should be read. The five options were about what consumed the project once it was underway. So this can’t tell you how many projects were sunk from the beginning by the schedule. Projects that were doomed by the deadline on day one were addressed by a different question (more about that in our next posts). What it does tell you is that, among the issues that crop up once a project is underway, the technology comes last – by some distance.
What the three real causes have in common
Waiting on another team. Requirements that moved. Data that’s worse than anyone had expected.
These aren’t specific technology problems. All three are often only discovered once the work has started. And all three sit in territory that a Head of IT controls: who owes what time to whom, who is allowed to change the scope, and how honest the organisation is willing to be about the state of its data.
That’s an interesting finding if the planning focused mainly on tooling.
What we take from it
We’re not going to pretend that 102 people at one conference settles anything – see the note at the end.
But it matches what we hear in scoping conversations. When we ask clients what was tricky about their past projects, almost nobody tells us that the software wasn’t up to scratch. If that’s right, the valuable questions to settle before a data project starts aren’t just about which platform. It’s who we’re dependent on, what we actually know about the state of the data, and who is permitted to change the scope once we begin.
Four more things in the data
Of the people who thought their deadline was unachievable, nine in ten said so – and half of those saw nothing change.
33% thought that the deadline was unrealistic.
55% spent half their project or more simply finding out what the data and systems actually looked like.
45% came in over budget. 7% came in under.
We’ll look at each of these properly over the next few posts.
About the survey
102 data practitioners – engineers, analysts and architects – completed a six-question anonymous survey on our stand at Data Decoded, Manchester, on 13 and 14 October 2026. No email address or identifying details were collected. Respondents were self-selected visitors to a single event, so this is not a representative sample of the UK market and we wouldn’t present it as one. What it is: 102 people who do this work for a living, describing it honestly with nothing at stake.
