Home / Directory / Performance and attribution

Performance and attribution: a buyer's guide

What we list in this category, how the field splits between long-established products and newer ones, and the questions worth pressing on before you commit.

Every entry below was recorded from a page we opened, and each carries the source we read. Where we could not establish something it is left blank rather than guessed. Read how this site works if you want the detail.

Long-established products

Recorded as established: in production for many years, usually with a deep reference base and broad coverage. Those are real advantages, and so is the fact that a fifteen-year horizon is easier to argue for a supplier that has already lasted one.

• Advent Geneva — Multi-asset portfolio and fund accounting system for alternative managers, with investor accounting and a full general ledger. source
• Aladdin — Enterprise risk, portfolio management and trading platform used across asset managers, pensions and insurers. source
• Axioma (attribution within SimCorp One) — Axioma portfolio construction, risk and attribution tools run inside the SimCorp One platform alongside SimCorp Dimension. source
• Axioma Risk — Cloud-native, API-first multi-asset portfolio risk system giving one consistent view of risk across a firm. source
• BarraOne — Multi-asset factor risk and return attribution across public and private assets, built on the Barra models. source
• Bloomberg PORT — Portfolio and risk analytics function on the Bloomberg Terminal, widely used for return attribution against benchmarks. source
• Cobalt — Private markets research, diligence, portfolio construction and monitoring platform; Hamilton Lane's own site lists Cobalt LP and iLevel as client portal options. source
• Eagle Performance — Performance measurement engine within the former Eagle Investment Systems suite, now marketed as BNY Eagle data and analytics solutions. source

Newer products

Recorded as modern or emerging. Newer suppliers often offer capability and pricing the incumbents do not, and they carry risks the incumbents do not: fewer references at your size, less operating history, and concentration if they are small. Both belong in the same comparison, on the same fields.

• Arcana — Portfolio intelligence for hedge funds combining a core factor model, thematic and intraday live risk, crowding and concentration analysis. source
• Clearwater Accounting & Reporting — Single-instance multi-tenant cloud platform that consolidates investment and fund accounting with reconciled IBOR and ABOR. source
• Essentia — Behavioral decision analytics for active portfolio managers, positioned as a complement to classical Brinson attribution by attributing return to manager decisions. source
• Factor Risk Lens — Factor risk decomposition for portfolio managers and traders, launched alongside Factor Analytics for CIOs and risk leaders. source
• Gyre Hub — Portfolio analytics platform covering portfolio management and construction, back testing, performance, risk, compliance, reconciliation, portfolio accounting, ESG and tax. source
• Kiski Platform — Risk decomposition, scenario analysis and stress testing aimed at independent managers, allocators and family offices. source
• Limina IMS — Cloud-native investment management system whose order management sits alongside portfolio management, compliance and operations; Arcesium announced the acquisition on 2 February 2026. source
• ODIN Platform and VICAP — Performance, attribution and multi-currency GIPS composite management, sold both as the enterprise ODIN platform and the subscription SAYS platform. source

What to ask

These come from the same question bank our request-for-proposal process uses, so a guide and an RFP never drift apart. The four sections below carry the most weight in that bank.

Functional fit

Whether the system does the job for YOUR instruments, markets and volumes — not for the demonstration portfolio. The questions separate what runs in production today from what is on the roadmap, because that distinction disappears in a sales meeting and reappears during implementation. It also asks what still has to be typed in by hand, which every system has and few volunteer.

• Which asset classes does the product support in production today?
• Is there a full audit trail of who changed what and when, and can it be exported?
• Does the system enforce segregation of duties between front and back office?
• Is maker-checker approval supported on material actions, and is it configurable by action type?
• Can a valuation be sourced and approved independently of the people who trade?

Security and resilience

Whether the vendor's security is independently attested or merely asserted. The questions ask for certifications you can read, penetration tests you can see the summary of, and recovery objectives that are written down. Incident history is asked directly, because a vendor that has handled a breach well is often a safer choice than one that has never been tested.

• Which independent attestations do you hold currently?
• Will you provide the most recent report under a non-disclosure agreement?
• Is data encrypted in transit and at rest?
• Is single sign-on supported, and is multi-factor authentication enforced?
• Can your own staff see our data, and is that access logged and reviewable by us?

Commercial terms

Total cost across the whole term, and what happens if you leave. Year one is asked separately from years two to five because year one hides where the money goes. The section asks explicitly what is NOT included, since the gap between the quote and the invoice lives there, and it asks about exit terms while you still have the leverage to negotiate them.

• How is the product priced?
• What is the total first-year cost for our stated requirement, including implementation?
• What is the total annual cost in years two through five?
• What is NOT included in that figure?
• Is there a contractual cap on annual price increases?

Artificial intelligence

If a model touches your data, on what terms. This section exists because the answers differ enormously between vendors and are rarely offered unprompted. It asks whether your data trains models serving other clients, whether you can opt out and keep full function, whether a human reviews output before it affects a book or a client report, and who is liable when a model is wrong. Skip it only if you genuinely do not care.

• Does the product use machine learning or generative AI anywhere in its normal operation?
• Is client data ever used to train models that serve other clients?
• Can we opt out of AI features entirely and keep full product function?
• Does our data leave your infrastructure to reach those providers?
• Is a human review step required before an AI output affects a book, a trade or a client report?

See all 47 products in Performance and attribution · Glossary