The best work happens in the spaces between disciplines

Ask anyone who's worked on a big digital programme where it hurt, and they almost never name a discipline - the strategists were sharp, the designers were good, the engineers knew their craft, the analysts had the numbers. What they name is the space between: the brief that lost something on its way from strategy to design, the design that met engineering reality too late, the data team measuring a product they'd never been in a room with. The work inside each box was fine, and the project still limped, because projects don't fail in the boxes. They fail in the gaps.

My job title says People and Operations, so I spend my days thinking about how teams are shaped rather than what they ship, and the longer I do it the more convinced I've become that the shape is not an administrative detail sitting underneath the work. An org chart decides where information is allowed to travel and who is in the room when a decision gets made, which means most of the gaps above were designed in months before anyone opened a brief. Handoffs aren't a communication problem you can meeting your way out of; they're a structural choice, and structures can be chosen differently.

What the gaps actually cost

Every handoff between silos charges three taxes. The first is translation: each discipline speaks its own dialect, and something is lost every time an idea crosses a border - the strategic 'why' arrives at design flattened into a feature list, the design intent arrives at engineering flattened into tickets, and by the third crossing the original thought needs a passport check to be recognised.

The second tax is time, and it's less about the waiting than about the queue-shaped thinking it produces: when your work goes over a wall, you optimise for what's defensible at the wall (complete documents, covered backsides, no loose ends) rather than for what the next person actually needs. The third tax is the quietest and the worst - accountability diffuses, because when four teams each own a quarter of the outcome, the outcome itself is owned by nobody, and everyone's metric can be green while the product underdelivers.

Four teams each owning a quarter of the outcome means the outcome is owned by nobody.

The seams are where the value hides

Turn the problem over and it becomes an opportunity, because the seams aren't just where failure lives - they're where the best decisions come from, when the disciplines are close enough to make them together. Strategy plus data is where you find out which customer problem is actually worth solving; design plus engineering is where you discover the elegant version that costs half as much to build; data plus design is where the funnel stops being a report and starts being a brief.

Our favourite decisions on any project are ones no single discipline could have reached alone - the engineer who suggests a simpler journey because they know what the API can do cheaply, the analyst who spots that the 'minor' edge case is a fifth of revenue, the designer who reframes the strategy deck's segmentation because they've actually watched the users in it. None of those moments survive a handoff process, because handoffs deliver conclusions, and these moments happen upstream of conclusions, in the messy middle where the thinking is still wet.

How we build for the seams

The structural answer is almost embarrassingly simple: one team per problem, all four disciplines in it, from the first workshop to the last release. Not a relay of departments with liaison meetings, but a pod that shares one brief, one metric and one standup - small enough that everyone knows what everyone else is wrestling with, and long-lived enough that the knowledge compounds instead of dispersing at handover.

The rituals follow from the shape. Discovery is attended by everyone (engineers hear customers firsthand rather than receiving personas by courier), reviews are cross-discipline by default (the data person critiques the design, the designer critiques the metric), and the shared number means nobody can win while the team loses. On the people side, we hire and grow for T-shapes - deep in one craft, conversational in the neighbours - because the seams only produce value when the people on either side can actually talk across them, and 'conversational' is a skill you can deliberately build.

None of this is free. Cross-discipline teams are harder to staff neatly, utilisation spreadsheets dislike them, and career ladders need more thought when people don't sit in discipline towers - we've felt every one of those costs, and we'd pay them again without hesitation, because the alternative costs more and hides it better.

The question worth asking your own structure

If your last three projects disappointed and the post-mortems all blamed 'communication', it might be worth entertaining the idea that nothing was wrong with the communicating and everything was wrong with the distance. Walls don't come down because people email across them more warmly, and the fix costs less than it looks: pick one problem, put all four disciplines on it together for its whole life, give them one number, and watch where the decisions start coming from. Our experience says they'll come from the seams - which is where they were waiting all along.