Replacing a Wealth Management Platform? Nine Mistakes That Derail the Project
Most wealth platform replacements don't fail on software. They fail on migration scope, booking-centre structure, suitability rules and advisor adoption. Here are nine costly mistakes and what to do about each one.
Mistakes one and two: scoping the replacement as a software swap
Replacing a wealth management platform at a bank or large wealth manager typically takes several years and often costs more than the original business case said. The overruns rarely come from the software. They come from decisions made, or skipped, in the first six months.
Mistake one: treating it as a like-for-like replacement. Teams list what the current system does and ask vendors to match it. That carries decades of workarounds into the new platform and skips the questions that drive cost. Which booking centres move? Which products stay on legacy custody? Which processes go to the vendor's operations service, and which stay in-house?
Instead: before you issue an RFP, write a target operating model. Cover entities, products, client segments, the split between in-house and outsourced work, and the systems that will sit around the core. Keep it short and decisive. If you can't say whether portfolio accounting is in scope, you aren't ready to pick a vendor.
Mistake two: running wealth management software selection off scripted demos. Every serious vendor can make a scripted demo look good, whether it's Avaloq, Temenos, FNZ, SS&C, InvestCloud, Objectway or a newer composable player. Scripted demos show the happy path, and nothing else.
Instead: once you're down to two or three vendors, run a paid, scenario-based proof of concept on anonymized data. Make the scenarios hard. A client booked in Geneva and served by an advisor in Dubai. A structured note going through a corporate action. A tiered fee schedule with household-level breakpoints. Budget roughly USD 50,000–250,000 per vendor. That's a typical 2026 range that varies widely by vendor and scope, so confirm before committing. It's small next to the cost of choosing wrong.
Mistakes three and four: underscoping position and history migration
Mistake three: assuming you migrate a snapshot of positions. Closing positions are the easy part. The hard parts are tax lots and cost basis held in more than one currency, daily valuation history for time-weighted and money-weighted performance, corporate action history, fee accruals partway through a billing period, and transaction records kept for regulatory reasons. Retention periods differ by jurisdiction and commonly run several years to a decade.
Instead: inventory the data and put every element into one of three buckets:
- Migrate: needed to operate, bill or report on the new platform.
- Archive with access: needed for audit, complaints or regulators, but not for daily work. A queryable archive is enough.
- Retire: no business or regulatory use. Get compliance to sign off.
Mistake four: leaving reconciliation until the final mock conversion. Take a hypothetical wealth manager with 60,000 accounts and 400,000 positions. Even a 0.5% break rate leaves 2,000 positions to investigate by hand on cutover weekend.
Instead: plan for several full mock migrations, at least three, with automated reconciliation. Units and cash should match exactly. Performance variance should sit within a tolerance you agree in advance, for example one basis point per period. Put migration tooling and reconciliation responsibilities in the vendor contract, not just the statement of work.
Mistakes five and six: booking centres and cross-border suitability rules
Mistake five: assuming "multi-entity" means multi-booking-centre. Many platforms can hold several legal entities. Far fewer keep them properly separated: separate books, holiday calendars, regulatory reports, and visibility rules that decide whether an advisor in Singapore can see a Zurich-booked client. Data residency makes this harder. Banking secrecy expectations in Switzerland, professional secrecy in Luxembourg, data protection laws in markets such as India, Saudi Arabia and Indonesia, and outsourcing notification rules in many jurisdictions can all dictate where data lives.
Instead: decide during selection whether you need one instance serving several entities or a separate instance per region. Test entity segregation in the proof of concept, not after signing.
Mistake six: hardcoding suitability rules for your home market. Each jurisdiction runs its own suitability and disclosure regime. Examples:
- EU: MiFID II suitability, appropriateness and cost disclosure.
- UK: the FCA's Consumer Duty.
- Switzerland: FinSA.
- Singapore and Hong Kong: MAS and SFC rules, including accredited and professional investor regimes.
- Dubai: DFSA rules in the DIFC.
- India: SEBI investment adviser regulations.
- US: Regulation Best Interest.
For cross-border business, the applicable rule can depend on where the client lives, where the account is booked and where the advisor sits, and these often point in different directions.
Instead: require a rules engine that compliance can configure along all three dimensions without vendor change requests. Ask who maintains the country rule content: the vendor, a third-party provider or you. Have local counsel in each jurisdiction validate the rule sets. These are legal judgments, and neither the vendor nor the project team should make them.
Mistake seven: treating the advisor front end as an afterthought
A common plan puts the core and back office first and promises advisor tools in "phase two." Then the budget runs out and phase two shrinks. Advisors end up re-keying data between the new core, their CRM and the spreadsheets they built around the old system. Adoption stalls, and the business case built on advisor productivity falls apart.
Instead: decide during selection whether the front end comes from your core vendor or from a specialist layer, and fund it in phase one either way. Specialist options include InvestCloud, Additiv, or a CRM-based front end such as Salesforce Financial Services Cloud or Microsoft Dynamics integrated with the core. A specialist layer offers more flexibility but adds integration work and a second vendor to manage. A single-vendor stack is simpler to run, but you're tied to that vendor's front-end roadmap.
Put five or six senior advisors from different booking centres on the selection panel and give them a real vote. Measure concrete tasks: building a rebalancing proposal, onboarding a client with a corporate structure, producing a client-ready review pack. If a shortlisted vendor's front-end capability exists only on its roadmap, treat that as a red flag.
Mistakes eight and nine: budgeting for licenses and cutting over in one weekend
Mistake eight: building the business case on licensing. License or SaaS fees are often a minority of the five-year cost. The table shows illustrative 2026 ranges for a bank or wealth manager with very roughly USD 20–80 billion in assets under administration. Pricing varies widely by region, contract structure and vendor, so confirm every figure with vendors.
| Cost driver | Indicative range (USD) | What pushes it up |
|---|---|---|
| Annual license or SaaS fee | 1.5M–8M per year | Pricing by AUM or account, number of entities, modules |
| Implementation and integration | 8M–40M one-off | Number of booking centres, custom interfaces, front-end scope |
| Data migration and reconciliation | 2M–10M one-off | Depth of history, quality of legacy data, number of mock runs |
| Dual running of legacy and new platforms | 0.5M–3M per quarter | Slipped decommission dates, waves that run long |
| Internal team and backfill | Often a significant share of external spend | Senior operations and compliance staff pulled off their day jobs |
Mistake nine: a big-bang cutover with no decommission date. Moving every entity in one weekend concentrates all the risk in one event. Leaving the legacy system running "just in case" keeps paying for two platforms.
Instead: migrate in waves by booking centre, starting with the simplest. Run each wave in parallel through at least one month-end, ideally a quarter-end. Agree a decommission date for each legacy system up front, and tie the vendor's milestone payments to it.
What to decide, and when, in a wealth platform replacement
Most of these mistakes are timing failures: the decision gets made after the cheap moment to make it has passed. Use this sequence as a guide for a typical multi-year replacement. Durations are indicative and will shift with scope.
| Phase | Typical timing | Decisions that must be closed |
|---|---|---|
| Target operating model | Months 0–4 | Entities in scope, in-house vs outsourced split, single vs multi-instance |
| Selection and proof of concept | Months 4–10 | Front-end approach, suitability rules engine, entity segregation tested |
| Contract | Months 9–12 | Migration responsibilities, decommission milestones, rule content ownership |
| Build and mock migrations | Months 12–24 | Data buckets signed off, reconciliation tolerances, wave order |
| Waves and decommissioning | Months 20–48 | Parallel run exit criteria, legacy shutdown per wave |
A rule for the steering committee: don't sign until the operating model is approved, entity segregation has been proven on your own data, and senior advisors have scored the front end. If any of the three is missing, hold the signature.
Common questions
How long does a wealth management platform replacement typically take at a bank?
For a bank or large wealth manager with several booking centres, roughly 30 to 48 months from operating-model design to final legacy decommissioning is common. A single-entity firm with clean data can sometimes finish much faster. Timelines grow mainly with the number of booking centres, the depth of history being migrated, and how much the front end is customized.
Should we pick a single-vendor stack or a best-of-breed front end on top of a core?
A single-vendor stack is easier to integrate and gives you one accountable party, which suits firms with limited integration capacity. A specialist front end gives you more control over the advisor and client experience, but it means more integration and two vendor relationships. Base the choice on which option your advisors scored higher in the proof of concept, not on architecture preferences alone.
How much transaction history should we migrate to the new platform?
Migrate what you need to run the business, bill clients and calculate performance. That often means full tax-lot detail plus several years of valuations and transactions. Older records needed for audit or regulators can usually go to a queryable archive. Retention requirements vary by jurisdiction, so have compliance and legal counsel confirm the cut-off for each booking centre.
Can one platform handle investment suitability rules across multiple countries?
Several leading platforms can, provided the rules engine can be configured by client residence, booking centre and advisor location, and someone keeps the country rule content current. Find out during selection whether that content comes from the vendor, a third-party provider or your own compliance team. Local counsel should validate the rule sets in each jurisdiction before go-live.
Written by the Equitystack Pro editorial team. Prices, features and rules change often - confirm details with the vendor or a licensed professional before you decide.