EN
Back to the archive

The encyclopedia · Finance & Accounting · Financial decision · 1998–2004

Bank Hapoalim built a portfolio optimizer for retail customers

Bank Hapoalim's Opti-Money built tailored asset-class portfolios, earning customers $244m above benchmark and the bank $31m.

Bank Hapoalim

the move

Bank Hapoalim, Israel's largest bank, needed a way for its relationship managers to give sound investment advice to individual customers, who differ in risk and situation and who had previously received inconsistent guidance.

Opti-Money is a decision-support system that takes customer- and market-specific inputs and solves a modified Markowitz-type nonlinear programming problem to produce an optimal tailor-made portfolio in asset classes, without making buy or sell recommendations.

Over five years it was used in 133,000 consultation sessions with 63,000 customers in 2002 alone; customers who used it earned 88% higher net income, advised customers earned $244 million a year above benchmark, and the bank's direct annual income exceeded $31 million.

why it works

  • Retail customers have different risks, so one-size advice is wrong.
  • Stocks and bonds are too fine-grained for a retail advisor to price.
  • Optimizing asset classes per customer makes advice tailored and consistent.
  • Measured: $244m a year above benchmark for customers, $31m for the bank.
the payoffSolve a portfolio once per customer, not per productinspired

what transfers

To give personalized advice to many customers, solve one optimization per customer over asset classes, instead of scaling product rules of thumb; consistency and tailoring are the value.

what came after

Opti-Money ran for years as a mature system and gave the bank a unique position in the Israeli banking sector, with income directly attributed to it exceeding $31 million in 2002; it was a Franz Edelman Award finalist because it scaled tailored private-bank-style advice to retail customers.

references

spotted an error? The archive wants to know.

same kind of clever