Getting Members to Stay: The Retention Problem Nobody Has Solved Yet

For CPF Providers, the best glidepath in the world cannot help a member who has already left.

Build the best glidepath in Singapore, stress-test it against fifty years of market history, and price it inside the 0.5% fee cap, and it will still fail the moment a member panics and exits at the bottom of a downturn. Most discussion of the CPF Lifecycle Investment Scheme treats this as a footnote. It deserves to be the second design problem, not an afterthought to the first.

Our previous post in this series covered the analytics infrastructure that makes a glidepath credible to the independent investment consultants reviewing provider submissions. This post is about a different audience inside the same pension provider organisation: the product leads, UX directors, and marketing heads who decide what a member actually sees, and when, across the life of their CPF investment journey.

The behaviour gap is well documented, and CPF will not be exempt

There is a substantial body of evidence on the gap between what an investment returns and what the average investor actually earns from it. Morningstar's Mind the Gap research has consistently found a meaningful annual shortfall between fund returns and the returns investors capture, driven almost entirely by the timing of buy and sell decisions rather than by the underlying investment. Decades of similar research, including DALBAR's long-running US studies, point to the same conclusion: badly timed decisions during periods of stress account for most of the gap, not poor portfolio construction.

There is no reason to expect CPF members will behave differently. If anything, the stakes are higher. CPF lifecycle products are designed to help members reach the Full Retirement Sum (FRS), currently SGD 220,400, or the Enhanced Retirement Sum (ERS) at SGD 440,800, by age 55. A member who exits during a downturn and re-enters later, or exits permanently into CPF's risk-free interest rate, has not just had a bad year but potentially converted a recoverable paper loss into a permanent shortfall against a hard, government-set target.

This is the part of the CPF scheme a glidepath alone cannot fix. The glidepath determines what happens to a member's allocation if they stay invested. Retention design determines whether they stay invested at all.

Four mechanisms, ranked by what they actually cost to build

Providers do not need to solve member psychology from first principles. There is a reasonably well-understood set of mechanisms for keeping people invested through volatility, and four stand out in a CPF context as both practical and aligned with the scheme's constraints.

Peer cohort transparency is the lightest touch and the easiest to ship. Showing a member how their position compares with others in the same age cohort, rather than an isolated number that looks alarming on its own, gives context a raw balance figure cannot. This is the same principle behind moving away from superficial gamification toward engagement built on genuine transparency: a 12% drawdown looks different when a member can see their cohort, on the same glidepath, has experienced something similar.

Hardship carve-outs with automatic re-entry address what is probably the single largest cause of permanent exits: members who genuinely need funds during a crisis, rather than members who are simply anxious. A defined hardship pathway, with automatic re-entry once the immediate need has passed, prevents a temporary liquidity problem from becoming a permanent shortfall.

Commitment fee rebates are the cleanest incentive available within the 0.5% fee cap. A small fee reduction tied to staying invested through a full market cycle directly rewards the behaviour the scheme needs, at minimal cost.

Guided pause or defer flexibility is the most structurally complex of the four, but arguably the most valuable. Rather than a binary choice between staying fully invested or exiting entirely, a guided pause lets a member temporarily halt new contributions to growth assets, or defer a scheduled de-risking step, without unwinding their position. It gives an anxious member somewhere to put their anxiety that is not “withdraw everything,” reflecting a wider shift Kidbrooke has written about previously: away from engagement mechanics that reward activity for its own sake, toward tools that support members through the moments that actually determine outcomes.

Note that none of these four mechanisms replaces good investment design. They need to be part of the product architecture from the outset, not bolted on once a provider notices members leaving.

Why this can’t be a UX afterthought

It is tempting to treat retention mechanisms as an interface problem to be solved late in development, once the analytics and the glidepath are finished. That sequencing gets the dependency backwards: every one of the four mechanisms above depends on infrastructure that must exist before the member-facing screen can be built.

Peer cohort transparency requires real-time access to anonymised cohort-level outcome data, generated from the same simulation engine that produces the glidepath, not a separate reporting layer. Hardship carve-outs need account architecture that supports partial liquidation and automatic re-investment, with projections that reflect the re-entry accurately rather than resetting to a generic starting point. Fee rebates need calculation logic that tracks holding periods at the individual member level without breaching the fee cap's reporting requirements. Guided pause functionality needs a contribution and rebalancing engine that supports a paused state distinct from an exit, with projected outcomes recalculated rather than simply frozen.

This is, in other words, an analytics and infrastructure problem wearing a UX costume. If a retention feature cannot draw on live, member-specific projection data it is merely decoration. KidbrookeONE's member-facing projection layer sits on top of the same Economic Scenario Generator and Monte Carlo simulation engine that powers glidepath design, so a peer cohort comparison or a guided pause recalculation draws on data and methodology the provider has already built and validated, rather than requiring a second, disconnected system.

The CPF Board has confirmed it will select only two to three providers for the scheme. With such a small field, providers are not competing on glidepath construction alone, since independent investment consultants will have already filtered for analytical credibility before any product reaches members. Among a small set of providers with broadly comparable methodology, member experience and retention design become a genuine point of differentiation.

The new scheme sits alongside CPF LIFE, the existing annuity scheme providing monthly payouts from age 65. CPF LIFE solves longevity risk after retirement; this lifecycle investment scheme solves the accumulation problem before that point. Retention design determines whether the accumulation phase produces a balance worth annuitising in the first place.

A glidepath cannot help a member who has already left

The investment design conversation around CPF has, understandably, focused on what providers can control with precision: asset allocation, rebalancing logic, fee structure. Retention design is harder to quantify and easier to defer. But a glidepath modelled to perform well across thousands of economic scenarios is only ever tested against the scenario where the member actually stays on it.

Providers who build retention thinking into the product from day one, drawing on the same analytics infrastructure that produces the glidepath rather than treating it as a separate workstream, will be solving a problem most submissions only address after launch, when fixing it costs far more than designing for it now.

If your team is thinking through what the member-facing layer of your CPF submission needs to do, get in touch.

Kidbrooke is a financial technology company providing unified investment and wealth analytics through KidbrookeONE, an API-first platform serving pension providers, asset managers, and wealth platforms.