Product portfolio management is the practice of running several products under one strategy and deciding what each of them is there to do. The discipline reached product management from corporate strategy. Bruce Henderson of the Boston Consulting Group published the first widely used version in a 1970 essay called The Product Portfolio. It sorted business units by market growth and market share, and it attached a standing instruction to each of the four quadrants. A product manager runs into all of this the day the organisation owns a second product. The leading and lagging indicators from the last page report on one product at a time, and a portfolio can fail in ways that no single product's numbers will ever show.
The sections below define what a portfolio holds. They then give the role each product plays and the measure that suits it. The three horizons model comes next, along with the criticism it has attracted. After that come the allocation ratios the research supports, and the page closes on the decision to wind a product down.
What a portfolio holds
A portfolio is the set of products an organisation funds from one pot and holds answerable to one strategy. Membership follows the money and the people. If two products are built by separate teams, paid for out of separate budgets, sold to separate buyers and reported separately, then they sit in two portfolios whatever the organisation chart says. If two products compete for the same engineers, they belong in one portfolio, because any decision about either of them is a decision about the other.
Managing them together lets an organisation move money and people as the evidence arrives. A product manager running one product improves that product. A product manager running the portfolio can hold a working product flat for a year and put the difference into one that will matter more in three years. No other level of the organisation can make that choice.
The role each product plays
Every product in a portfolio is there for a stated reason, and the reason decides the measure. A product defending the revenue that pays for the rest is judged on retention and margin. A product testing whether a market exists at all is judged on what the team has found out for the money it spent. Five roles cover almost every portfolio.
| Role | What the product is for | The measure that suits it |
|---|---|---|
| Grow | Taking share in the segment the strategy chose | New customers, share of the segment, rate of activation |
| Defend | Holding the revenue and the customers the rest of the portfolio spends | Retention, margin, cost to serve, win rate against the named competitor |
| Harvest | Funding the rest of the portfolio for as long as demand lasts | Cash generated per engineer, gross margin, support cost per account |
| Explore | Finding out whether a market and a willingness to pay exist | Assumptions retired this quarter, cost of the evidence bought |
| Wind down | Releasing money and people into the other four roles | Customers migrated, cost removed, closing date met |
Judging all five roles on growth is the common error, and it does damage in two directions at once. A harvest product measured on growth attracts investment it will never repay, because the market it sits in stopped growing before the investment was approved. An explore product measured on revenue in its first year reports a failure. At that stage the only honest measure is what the team now knows and what that knowledge cost.
The three horizons model
Portfolio roles say what each product is for today. A separate question is where the next generation of products comes from. Most organisations answer it with the three horizons model, published by Mehrdad Baghai, Stephen Coley and David White of McKinsey in their 1999 book The Alchemy of Growth. The model sorts growth work into three horizons and asks an organisation to fund all three in the same year.
- Horizon one is the core business, which earns most of today's revenue and takes in the steady stream of small improvements.
- Horizon two is the set of emerging businesses that already earn something and still need investment to reach scale.
- Horizon three is the set of options on businesses nobody sells yet, and it holds research, pilots and small stakes in other companies.
The three horizons argue for funding all three at the same time. An organisation funding only horizon one reports excellent numbers for as long as the core lasts. When the core weakens there is nothing standing behind it, so the revenue cliff arrives on a date somebody could have worked out years earlier.
The pale part of each band is money going out and the solid part is money coming back. Every band starts at the same edge, because an organisation that waits for the core to weaken before funding horizon three has already spent the years horizon three needs.
The fair criticism of the three horizons model
Funding all three at the same time is the part of the model that survives. The timings attached to it have been attacked hard. Steve Blank made the criticism worth taking seriously on 8 January 2019, when he argued that the arithmetic of time inside the model has stopped holding. His claim is that horizon three ideas can now be delivered as fast as horizon one ideas. A challenger puts existing technology together into a new business model in months, and it carries none of the legacy an incumbent has to work around.
Two things follow for a portfolio. An organisation that treats horizon three as work for later years will meet a competitor who treated the same idea as this quarter's work. And once horizon three is read as a date far away, it becomes the place an organisation files the ideas it has no intention of funding. That is how a portfolio review can name a dozen future businesses and fund none of them.
Blank's own remedies are about speed. An incumbent facing a fast challenger can pay outside parties to build the product, buy the company that has already built it, copy the model quickly or simply move faster than the challenger did. All four of those are portfolio decisions about where the money goes this year.
Allocation across the portfolio
Decisions about where the money goes each year have published numbers behind them. Bansi Nagji and Geoff Tuff put them on the record in Harvard Business Review in May 2012, after studying how companies spread innovation investment across the core, the adjacent and the new. Their finding was that outperforming companies put roughly seventy per cent of innovation resource into the core, twenty per cent into adjacent work and ten per cent into transformational work. Returns ran the other way, with around seventy per cent of the return coming from the transformational tenth.
The ratio is a starting position for an argument and no more than that. Nagji and Tuff said the right balance differs by industry and by company. The useful part of the number is that it gives the transformational tenth a name and a figure. Somebody can defend a line item in a budget meeting. An unnamed intention disappears in the first round of cuts, and nobody even records that it happened.
Winding a product down
A round of cuts is one way a portfolio loses a product. A deliberate wind down is the other, and it is the role an organisation almost never carries out. Clayton Christensen explained the mechanism in 1997 in The Innovator's Dilemma. Resource allocation inside a healthy company routes money towards the customers it already has and the margins it already earns. The product that would replace the core therefore loses the internal argument on every measure the company trusts. The same machinery keeps weak products alive, because nothing inside it ever produces a decision to stop.
A wind down needs four things written down before it will actually happen.
- A closing date.
- A migration route for the customers who remain.
- One named owner with the authority to hold the date.
- An account of where the engineers go on the day after.
The account of where the engineers go decides whether the other three actually happen, because a team with no next assignment will keep finding reasons for the product to continue.
All five portfolio roles assume an organisation can move money and people between products once the evidence changes. If a year's money was committed against a fixed scope twelve months ago, a portfolio role is only a label on a slide. Nobody can rebalance anything until the next cycle opens. How the money arrives therefore decides what a portfolio can do.