Lluis Gasso — Fractional Head of Analytics

The 9am email

It’s 8:55am. The CEO has emailed: “Why is ROAS down?” You open the dashboard. Then a second one. Then a third. Three platforms, three numbers, three definitions. By 6pm you’re more confused than you were at 9am.

Most marketing leaders at $50M+ DTC eCom companies have lived a version of that morning. The fix that’s being sold to you almost never addresses the actual problem.

You’ve already tried the fixes

Manual analyses that arrive late. Attribution platforms whose numbers nobody trusts. MMMs obsolete by the time they land. Vendors pitching AI agents on top of the same broken foundation. Big consultancies selling six-month frameworks that look great in the deck and never survive contact with the warehouse.

None of it works because none of it addresses the actual problem.

What I learned the hard way

I’ve spent the last 14+ years living through every analytics technology cycle. Tag management. Data layer standards. Cloud data warehouses. Reverse ETL. Customer data platforms. The transformation layer. Now AI agents on top of all of it.

Each cycle was sold the same way: this is the layer that finally makes your data trustworthy. Each cycle hit the same wall, the wall that has nothing to do with the technology.

Every technology can produce reliable numbers. None of them can make your departments agree on the same definition of revenue. None of them can stop a data engineer from building a duplicate table because they can’t see what depends on the original. None of them can reconcile a CMO who calculates ROAS on Gross Revenue with a CFO who reports Net.

The pattern was hiding in plain sight. The technology cycle wasn’t the cause of the problem. It wasn’t the solution either. The problem lived between people, and the technology kept being asked to solve it.

The reframe

Most data problems aren’t accuracy problems. They’re trust problems between teams.

Net Revenue is the cleanest example. Marketing wants Gross. The bigger number, the faster number, the one that reflects better. Finance wants Net. After discounts, with taxes, without shipping costs. Operations wants something in between. Nobody has explicitly disagreed. Nobody has lost a meeting over it. So the same dimension drifts toward whoever’s doing the report at the time. A marketing report uses Gross. A finance report uses Net. The CEO sees both numbers and can’t tell whether sales are up or down.

That’s a governance problem disguised as a data-quality problem. Buying a new tool doesn’t fix it. Visible lineage, smaller iterations, and difficult conversations brought to a foundation everyone can see does fix it.

What I do now

I’m a Fractional Head of Analytics for $50M+ DTC eCom. Two-week DAR loops anchored to the Horizon. Diagnose what’s actually broken, with evidence. Act on the smallest unit that ships value. Reflect with stakeholders before the next loop. The Horizon is the success definition stakeholders sign off on at the start. It does not move.

Most engagements start with The First Loop: 4 weeks, fixed scope, fixed price. One trusted artifact shipped end-to-end, the Horizon defined, and a written proposal for an ongoing engagement, or a clean exit with real value in hand.

What I won’t do

I won’t sell you tools I’m a partner of. There are no preferred partners. No reseller margins. No incentive to recommend a stack I get paid on.

I won’t write a six-month framework. The math doesn’t work. Your warehouse will outlast the deck. We ship in two-week increments because that’s the only cadence that survives contact with reality.

I won’t deploy AI on broken governance. Garbage in, garbage out, on steroids. The work earns the right to deploy AI. AI does not earn the right to skip the work.

I won’t take engagements at companies under $10M revenue. Different problem. Different cost of advice. Earlier-stage companies need someone who can do every job, not a Fractional Head of anything.

Recent work

A $100M+ DTC eCom asked me to build dashboards. After a month inside the stack I told them their dashboards weren’t the problem. Two months later: a 9am dashboard the whole org trusts, definitions every department signed off on, 66% reduction in data processing costs. Two months. No new tool.

Read the case study →

The track record

Director of Engineering at Honeylove. Founder of Data Crafts, a data engineering consultancy. Senior Analytics Consultant at Search Discovery designing measurement frameworks deployed across hundreds of enterprise sites. Implementations across J&J (600+ global sites), Equifax, LaunchLife, and SKY TV. Earlier work at Havas Media UK on Adobe Analytics and tag management for global brands.

Where I write

I publish field reports from inside DTC data engagements at Ground Truth. One short note a month. What worked, what didn’t, what I changed my mind about. No fluff, no AI hype.

A personal note

I came up through implementation. Tagging, data layers, ETLs. The kind of work where you can see whether your number is right or not within an hour. I never lost the bias for that, the preference for shipping a small thing that works over describing a large thing that might.

The thing I believe most strongly about the analytics industry: most of the work that’s sold as strategic is alignment work in disguise, and most of the work that’s sold as technical is governance work in disguise. The unglamorous middle, naming a number so the same answer comes back regardless of who’s asking, is where the leverage actually lives. That’s the work I do.