Marketing Attribution for Asset Managers: Why It Breaks and What to Measure

Distribution & GTM · The Last Mile

Every asset manager I meet wants to prove marketing moves flow. Almost none can, because the flow ends inside a bank shelf they do not own. Here is why attribution breaks, and the four layers you can actually measure.

Victor Roodhorst · roodhorst.eu · August 2026 · 7 min read

Every quarter, somewhere in a European asset manager’s office, someone asks the same question: show me what marketing did. In my years running a £10M European acquisition budget at IG Group, I could answer that question. I could point at the click, the open account, the first deposit. In institutional asset management, where I sit now, the honest answer is different. The click stops at the gatekeeper.

The structural problem

Why attribution models break in asset management

The standard attribution toolkit was built for journeys that end in a conversion you can see. Asset management does not have one. Three structural reasons explain the gap, and none of them is a measurement-tool problem.

The journey ends inside a shelf you do not own. Banks hold roughly a 57% share of EU retail fund distribution, according to EFAMA. The mandate decision happens on a buy list, in a due-diligence read, in a preferred-partner meeting. Your last click is a whitepaper download, not the decision. The Financial Communications Society calls this the last mile: the part of the journey, through intermediaries, that most marketing measurement simply cannot see.

The cycle is long and the conversion is not a click. Institutional mandates run for years. The conversion event is a request for proposal, a consultant recommendation, a board approval. An attribution window of thirty days assigns nothing, because nothing happened inside it that you could call a purchase.

Marketing and distribution keep two scoreboards. Web dashboards live in marketing, coverage and pipeline live in the CRM, and nobody has joined them on one key. Until they do, no model can connect effort to flow. This is the exact failure the Distribution Engine’s unified data layer exists to fix.

Industry observers have stopped calling this optional. VettaFi put it plainly this year: accurate attribution is no longer optional, because firms without an attribution strategy face inefficient spend and a clouded view of true performance. Agreed. The problem is that most of the advice aimed at the industry starts at the model selection screen, and the model is not where asset management breaks.

The frame

Banks own the last mile

Before choosing an attribution model, be honest about where the flow actually lives. In Europe, most of it sits behind a bank. EFAMA’s data puts banks at a 57% share of retail fund distribution, and in some markets the concentration is sharper still: in Belgium, banks distribute well over EUR 240 billion in funds issued outside their own scope, on an open-architecture basis. The shelf is the conversion event. The three channels that carry European fund flow all route through gatekeepers: the bank’s proprietary channel, the open-architecture platform, and only a small direct layer where the investor is reachable at all.

That is why asset-management attribution is not a web-analytics problem. It is a distribution-visibility problem. You can measure the click, and the click will be real, and the click will not be the decision. The decision happens on a shelf you do not own, in a meeting you are not in.

The diff

What the last-click model actually tells you

The difference is not a data problem. It is a model problem. You are measuring the only event you can see and calling it the decision, and the model faithfully reports the click while missing the shelf. Fix the model choice and the click still is not the decision. You have to fix what you can see first.

The ledger

The four layers you can actually measure

You cannot see the shelf decision directly. You can instrument everything around it. Four layers, in the order they pay back.

Layer 1. Your owned web: intent, not clicks

Which funds get read, which PDFs get downloaded, which webinars the account team attends. This is the one layer marketing fully controls, and it is best used to score accounts, not to claim credit. An account that reads a fund’s literature is an intent signal, and intent signals feed coverage, not last-click dashboards.

Layer 2. Sales and coverage data: the CRM ledger

Meetings, proposals, stage movement, buy-list reviews. Attribution is only as good as the client master it sits on; most attribution failures are data-model failures before they are analytics failures. If the CRM cannot join an account to its interactions, no model can either.

Layer 3. The distribution layer you can see

Platform data where visible, buy-list outcomes, flow data for the instruments you control. Even partial sight beats a black box. It tells you which channel actually moved, and that is the event worth attributing to.

Layer 4. Brand and recall: the leading indicators

Brand surveys and AI share of voice measure the recommendation moment, before traffic and before pipeline. In a gatekept market, where the buyer asks a machine for a shortlist before calling anyone, recall inside the AI answer is the earliest signal you can buy. The name is part of that recall: European fund names are nine words long on average, and getting longer.

The model

Which attribution model fits your channel mix

Once the ledger exists, the model choice becomes straightforward. VettaFi’s framing is the useful one: the model follows your data maturity and your channel mix, not the other way around.

The operator read

Start with the ledger, not the tool

Do not buy an attribution platform first. The industry pattern is tool-first, data-second, and it fails, because the tool has nothing to join. Start with the unified data layer: one definition of a tier-A client in one system, web and CRM joined on one key. That is layer three of the Distribution Engine, and it is the layer that quietly caps every other.

Then pick the model for the channel mix, and instrument the leading indicators quarterly. The observatory exists for exactly that reason: it measures whether the machine names your firm at the moment of recommendation, which in a gatekept market is the earliest signal of whether the shelf will ever see you.

Frequently asked

What is marketing attribution in asset management?

Marketing attribution is the attempt to connect marketing activity to revenue outcomes. In asset management it is unusually hard because flow runs through intermediaries, the sales cycle is long, and the conversion event, a place on a bank's shelf or a mandate award, is rarely visible to the marketer.

Why is marketing attribution so hard for asset managers?

Three structural reasons: the journey ends inside a distributor's shelf, the cycle is measured in years, and marketing and distribution keep separate data systems. None of them is solved by a better model; they are solved by seeing the flow first.

What attribution model should an asset manager use?

The one that fits your data maturity and channel mix. Single-touch suits short retail journeys, multi-touch suits mixed journeys, data-driven models suit mature unified stacks, and media mix or geo lift tests suit shelf and coverage decisions where clicks do not exist.

Can asset managers measure distribution through intermediaries?

Partially, and that partial view is the whole game. Platform data, buy-list outcomes, and flow data for the instruments you control tell you which channel moved. Pair that with CRM coverage data and leading indicators like AI share of voice, and you have a usable picture of the last mile.

Victor Roodhorst publishes the AI Share of Voice Index for European Asset Management and runs roodhorst.eu, where he publishes original studies on how European financial services actually markets. He works from the regulated-operator seat, not the agency one.

For the longer treatment of the same argument, read the 2026 marketing playbook.