Ansoff Matrix Example: The Growth Move That Isn't on the Grid

The Ansoff matrix is usually taught as a menu. Four cells, pick one, each with a risk rating attached. It is a good menu, and the site has worked it on Apple, Nvidia and Tesla.

This is a Decision Replay, so the question is different: take a real, dated, publicly announced growth decision with a known outcome, run the matrix on it, and see whether the matrix could have produced the answer.

Here it could not — and not because the answer was hard. Because the answer is not on the grid.

The decision

On 5 February 2014, Basecamp's tenth birthday, Jason Fried announced that 37signals would sell, shut down or spin off every product it made except Basecamp, and rename the entire company after the survivor. Highrise (CRM), Campfire (chat) and Backpack went; Ta-da List and Writeboard had already been in maintenance mode since 2012.

His stated reason at the time:

"We do our best work when we're all focused on one thing."

Eight years later he described the step himself, in a post announcing that they were undoing the name:

"About 8 years ago in 2014 we decided to go all-in and focus on just one product moving forward. That product was Basecamp, our biggest hit. To eliminate any doubt about our newfound focus, we renamed the entire company."

What was on the table in January 2014

Core product	Basecamp — ten years old, market-leading in SMB project management
Other live products	Highrise (CRM), Campfire (chat), Backpack
Already sunset	Ta-da List, Writeboard (maintenance mode since 2012)
Market	Small businesses and teams — broadly the same market for every product
Ownership	Private, profitable, no outside growth mandate
Constraint Fried names	Attention. Not capital, not market size

Note the third row, because it determines where everything sits on the grid: Highrise, Campfire and Backpack were new products sold into the existing market. That is textbook product development. 37signals in early 2014 was a company already occupying two Ansoff cells at once.

Running the matrix

Cell	What it would have meant in 2014	Ansoff's conventional read
Market penetration — Basecamp, existing market	Sell more Basecamp to more small teams	Lowest risk. Do this first.
Product development — new products, existing market	Keep building Highrise, Campfire, Backpack	Already in flight
Market development — Basecamp, new markets	Take Basecamp upmarket to enterprise, or into new geographies	Never seriously pursued
Diversification — new products, new markets	Something unrelated, for someone else	Never on the table

Four options. Now compare them against what actually happened.

None of the four is "sell Highrise, shut Campfire, and put the entire company on Basecamp."

The closest cell is market penetration, and it is not the same thing. Penetration says sell more of the existing product. It is silent on whether you keep the other three. A company can pursue penetration on Basecamp and keep Highrise running — indeed that is exactly what 37signals had been doing. The 2014 decision was not "do penetration." It was "stop doing product development, and free the attention."

That move has no cell. Look at the axes: products — existing or new; markets — existing or new. Neither axis has a value for fewer. Every one of the four cells is an addition. The matrix is structurally incapable of returning subtraction as an answer, because subtraction is not a coordinate on it.

The Decision Replay

What the framework says	What they actually did	What actually happened
Four growth options, ranked by risk. The lowest-risk recommendation is market penetration — push Basecamp harder into the SMB market — while product development continues in parallel. Nothing in the output suggests removing anything.	On 5 Feb 2014 sold, sunset or spun off every product but Basecamp, and renamed the company Basecamp to make the commitment unambiguous.	Basecamp remained the core business through the focused years. In 2018 they began work on new email ideas, launched HEY in early 2020, and on 3 May 2022 Fried renamed the company back to 37signals — "We aren't a one product company anymore. We're two. Basecamp and HEY."

The sequence, in Ansoff's own vocabulary: penetration → (subtraction) → penetration → product development.

Three of those four steps are on the grid. The one that isn't is the one that made the last one possible — and that is the claim worth testing rather than asserting, so here is Fried's own framing of the difference between 2014 and 2022. In 2014 the company had to leave things behind to focus. In 2022 he describes a company that can "do many things at once without leaving anything behind."

That is a statement about capacity, and capacity is what the focused years bought.

The Fifth Cell

The Fifth Cell — every quadrant of the Ansoff matrix is an addition, because neither of its axes has a value for "fewer." Before using the matrix to choose a growth direction, ask the subtraction question separately: what would we remove, and what would that free? The matrix cannot raise this question, and it will never appear in its output, so it has to be asked outside the tool.

Two things make this more than a pedantic observation.

First, Ansoff is used as a menu, and menus imply completeness. Nobody opens a 2×2 thinking "these are four of the five things I could do." The visual form of a completed grid is an argument that the option space has been covered. It has not.

Second, subtraction is frequently the enabler of the cells that are on the grid. The 37signals path is the clean illustration: the product-development move that produced HEY came after eight years in which the company deliberately did less. A 2014 planning session that ran Ansoff and picked from its four options would have kept Highrise and Campfire alive, kept the attention split, and arrived at 2020 as a weaker company — while having followed the framework correctly.

The practical version is one line added to the front of any Ansoff session:

1. Before opening the matrix: what could we stop doing, and what would that free up?
2. Then open the matrix, and read its four cells as growth options rather than as the full option set.
3. Check whether the answer to (1) is a precondition for the cell you picked in (2). If the cell you want requires attention you do not have, the fifth cell comes first.

Where this misleads

Ansoff never claimed to cover shrinkage, and criticising a growth matrix for not modelling divestment is partly unfair. Igor Ansoff built it in 1957 to answer how do we grow, and it answers that. The failure mode is in the usage, not the design — but the usage is near-universal, and a tool that is always used as a complete option menu should be evaluated as one.

37signals is close to a best case for subtraction, and most companies are not. Private, profitable, no investors, no board demanding a growth rate, and two founders with enough public standing to absorb the reputational cost of killing shipped products. A venture-backed company at the same juncture usually cannot choose the fifth cell at all — not because it is wrong, but because the capital structure forbids it. Read the finding as "the matrix cannot express this option," not as "you should take it."

Subtraction also kills companies, and we mostly hear the survivors. For every focus story that worked there are firms that cut the line that would have carried them. Nothing here says removal is safe; the claim is only that it is an option the grid cannot show you.

We do not actually know that the subtraction caused the later capacity. Eight years is long, HEY's success has many parents, and Fried is an interested narrator of his own company's history. The sequence is documented and the causal link is his account of it, not a measured result.

Now run yours

1. Ask the subtraction question first, in writing, before anyone draws the grid. What would we stop, and what would it free — people, attention, support load, roadmap?
2. Then run the matrix on the four growth cells, honestly labelling which ones you are already in. Most companies occupy two and only notice one.
3. Test the pairing. If the cell you want needs attention that only subtraction can free, sequence it: fifth cell first, growth cell second.

Want to map your own growth options? Framework for iPhone & iPad ships an Ansoff worksheet with AI assistance alongside 100+ other frameworks.

Want to go deeper

• Ansoff Matrix — the four cells, the risk ordering, and how to place an option
• Ansoff matrix examples — the method run across several companies
• Apple Ansoff Matrix (2026) — a company occupying three cells at once
• Pre-mortem Example: The Failure a Premortem Cannot Catch — another Decision Replay where the framework's structure is the finding
• Pros and Cons Example: 9 Pros, 5 Cons, and the One That Decided It — a replay where the framework is right and still does no work

Cover photo: Andrea De Santis on Unsplash.