Product Portfolio

With one product, nobody needed the word. Ten people were available and all ten worked on that product, and whether it was going well came down to a single number.

The second and third products changed that. Each team argued their product was the urgent one, and each of them was usually right. A new product needs people now or it never gets off the ground, and an old one carries half the revenue, so taking hands off it puts that half at risk. The problem was what to do when three correct arguments collide and nothing exists to choose between them.

Quarterly planning meetings resolved that by volume instead. Three leads each ask for two more engineers, and the evidence each brings is the growth rate of their own product. Between a small product growing 40 percent and a large one growing 8 percent, that number alone does not say where the engineers belong. Without some view of the whole company on one board, the call gets made by whoever is most persuasive in the room that afternoon.

A portfolio works by assigning each product a distinct role, then attaching a different yardstick and a different share of resources to each role. There are three steps.

1. List every product and give each one a role. Roles usually fall into four groups: the cash product carrying most of the revenue, the growth product being scaled right now, the experimental product that may or may not work, and the defensive product that keeps customers from leaving. Some teams assign roles using Three Horizons, a thinking tool that splits a business into what earns today, what earns in two or three years, and what lies beyond that.
2. Set separate success criteria per role. The cash product is judged on revenue and margin, and a low growth rate is not a problem there. The growth product is judged on user growth, and losing money is acceptable. The experimental product is judged on whether the hypothesis held, not on revenue. Apply one yardstick to all four and the experimental product comes last every time, which removes the reason to ever start anything new.
3. Write the resource split down as numbers and revisit it quarterly. Something like 70 percent of people's time on cash and growth, 20 percent on defense, 10 percent on experiments. Leave it unwritten and every urgent situation pulls people off the experimental side first, until six months later that product exists in name only. The split does not have to be precise, but it has to be written somewhere a person can point at in the meeting where someone asks to borrow the experimental team's engineer for a fortnight.

Putting everything on one board also surfaces things invisible product by product. If all four products chase the same customers with the same features, that isn't a portfolio, it's one product four times over. If they have no relationship at all, the case for selling them together — or housing them in one company — gets thin. What to look for is whether products pass customers to each other, and whether they deepen the same moat, the structural advantage that gets harder for competitors to cross over time.

The board earns its keep in quarterly and annual resource planning. Four uses come up most often.

1. Splitting people and money — compare by role and share rather than by growth rate. This is where a decision like pulling one engineer off the cash product growing 8 percent and moving them to the growth product gets made. The board makes the trade legible: the cash product gives up a little speed it was never being judged on, and the growth product gains the thing it was short of.
2. Deciding what to shut down — a product whose role has gone empty becomes a candidate. Small revenue, no help to the other products, and the experiment already concluded. Without the board, the experimental product tends to get cut first purely for losing money.
3. Filling gaps — drawing the board exposes roles nobody occupies. Then comes the build-buy-partner question of whether to build into that gap, acquire someone, or partner.
4. Explaining outward — the first slide investors and board members see is usually this board. A list of products says less than roles and shares, which fit the whole direction of the company onto one page. Product positioning, the place each product claims in a customer's head, gets attached alongside.

Drawing the board is not the same as deciding. Product strategy — the choice of what to do and what to leave alone — comes first, and the portfolio is that choice written out product by product.

"What's this product's role? If it's cash, stop judging it on growth. If it's growth, stop treating the losses as a problem."

Someone is holding a product to the wrong yardstick.

"The 10 percent for experiments stays untouched. Pull it again and we have nothing to ship next year."

A commitment not to raid the experimental side when something urgent lands.

"There's an empty defensive slot on the board. Buying is faster than building here."

A gap was found and acquisition is being proposed over building into it.

Pagedeck still has one product, so there is no board to draw at the product level. Grace draws it across three revenue lines instead: the free plan, paid team plans, and enterprise contracts. Enterprise is the cash role, paid team conversion is the growth role, and the free plan plays defense by bringing people in. When Mia builds a quarterly plan she splits people's time across the three lines first and picks the work items second.

The most common mistake is reading more products as a stronger portfolio. The picture of one product wobbling while the others hold it up sounds reasonable. But each product carries its own headcount and maintenance cost, and running five of them with the same 45 people leaves all five half-finished. Adding products whose roles overlap is worse still. Sell two similar things to the same customers and they split each other's revenue while sales has to guess which one to recommend on every call, which shows up as a slower sales cycle long before it shows up in revenue. What makes a portfolio hold up is non-overlapping roles, not a larger count.

Expecting every product to grow is the other one. Growth feels unambiguously good, so the assumption slips in easily. But the cash product's job is to earn steadily and cover the growth product's losses, so attaching a growth target to it pushes someone into raising prices or rushing features that unsettle existing customers. Six months later the growth rate is up three points and churn has doubled. Different roles need different targets.