CloudAttribution has spent thirteen years working with large global asset managers — the kind with multi‑asset desks, layered benchmarks, external mandates and clients who ask hard questions in writing. The platform reviews we run with those firms have turned out to be the most useful thing we do, and not only for them. Held often enough, across enough firms, they stop being individual conversations and start being a map.
What the map shows is that attribution problems are remarkably consistent, and that they are operational and structural rather than quantitative. The models are rarely the thing that is broken.
Six problems, in almost every review
1. Reporting is no longer enough
Clients have become more sophisticated faster than the systems serving them. Teams are asked for cuts of the data their platform was never designed to produce, and the gap is filled by hand. The cost is not only the hours; it is the lag, and the quiet accumulation of risk in a process nobody has signed off.
2. The narrative and the numbers disagree
A manager explains in a client meeting what they were trying to do. The report shows allocation and selection by sector. Both are true and neither explains the other. As the questions get sharper, that gap becomes progressively harder to talk around, and confidence erodes on both sides of the table.
3. Spreadsheets hide the real problem
Manual reconstruction is treated as a workaround. In practice it is a disguise. When attribution is rebuilt in Excel, different versions tell slightly different stories, and internal confidence in the numbers goes first — before any client notices.
4. Nobody can separate mandated from discretionary
Multi‑asset teams in particular struggle to show what they actually decided, because the portfolio contains decisions that were handed to them. Without that separation, a team cannot evidence its own contribution, and accountability becomes a matter of assertion.
5. Look‑through and FX break first
Layered structures lose lineage to the underlying holdings. Currency attribution, often a material driver of return, is where most systems are found out. A great many products claim to handle both; considerably fewer survive contact with a real multi‑asset portfolio.
6. The system starts shaping the portfolio
This is the one that should worry you, and it comes up more often than anyone would like to admit.
Teams adapt their investment process to fit what their attribution system can measure — choosing a hedging structure because the alternative cannot be decomposed, or aligning to a benchmark the system can handle rather than the one that matches the mandate.
At that point the tail is wagging the dog. A measurement system has quietly become a constraint on investment decisions, and because the change happens gradually and for sensible local reasons, it is almost never visible as a decision at all.
What those firms did about it
The effective responses had something in common: they started from the investment process rather than from a feature list. Firms redefined what attribution had to deliver before asking who could deliver it.
- Transparency first. Complete visibility into what each manager did, how each strategy performed, and how individual decisions contributed. Evidence in place of inference.
- Accountability made measurable. Mandated and discretionary components separated explicitly, so a team can demonstrate ownership of what it controls.
- The narrative and the numbers reconciled. Strategy and trade‑level tagging drawn from the firm's own sources, so a position can be tracked from entry to exit and measured as the decision it was, rather than as whichever sector it landed in.
- Attribution used as a diagnostic. Granular enough to find strengths and weaknesses that were previously invisible, which turns reporting into a feedback loop.
- Operational friction removed. One consistent framework serving portfolio managers, performance, client teams and management, instead of four partial views that disagree at the edges.
- Wholesale replacement avoided. Almost nobody who looked seriously at decommissioning their platform concluded it was worth the disruption. The firms that got somewhere added a capability alongside what they had.
Why smaller firms were locked out
None of that is secret, and none of it is specific to size. A family office running thirty portfolios has the same six problems as a firm running three thousand. What it has not had is a route to the solution.
The barrier was never the mathematics. It was that attribution of this quality arrived as an enterprise programme: a multi‑year implementation, a mapping project, a professional services line, and per‑seat licensing that made every additional reader a cost to be justified. Priced that way, the capability is unreachable below a certain scale — and above it, rationed.
What Attribution Express changes, and what it does not
Attribution Express is the same thinking, delivered differently. It runs in your own Azure tenancy, on your transactions, your prices, your classifications and your analytics, under the licences you already hold. Nothing is redistributed and nothing leaves your systems.
What has not changed is the calculation. The decomposition a thirty‑portfolio firm gets is the decomposition we built for firms with a thousand times the assets, because there was never a good reason for it to be anything else. The difference is in how it is implemented, how it is priced, and who is allowed to open it.
If any of this is familiar
The six problems above are not a sales list. They are what firms told us, repeatedly, when we asked them what was actually going wrong. If more than two of them describe your month end, the constraint is probably not your team.