The BCG Matrix and Why It Falls Short for Modern Products
February 15, 2024
The BCG Matrix was created by Bruce Henderson in the early 1970s and became one of the most widely used tools in business strategy. By the late 70s, half of all Fortune 500 companies were using it. Harvard Business Review eventually named it one of five charts that "changed the world."
The idea is simple. You plot your products on a two-by-two grid. One axis is market share (how dominant you are relative to competitors). The other is market growth rate (how fast the overall market is expanding). That gives you four quadrants:
- Stars: High market share, high growth. Invest here.
- Cash Cows: High market share, low growth. Milk for profits.
- Question Marks: Low market share, high growth. Decide whether to invest or cut.
- Dogs: Low market share, low growth. Usually divest.
The logic follows from there. Use profits from your Cash Cows to fund your Stars and promising Question Marks. Cut your losses on Dogs. It's a resource allocation framework, and for a company managing a portfolio of consumer goods or physical product lines, it still does what it promises.
The problem is that most product teams today aren't managing detergent SKUs.
Where It Breaks Down
The BCG Matrix assumes you're competing in a market that already exists, with a growth rate you can measure and a market share you can calculate. That's a reasonable assumption for mature industries. It doesn't describe what most software and platform companies are doing.
Consider AWS. When Amazon launched it in 2006, cloud computing wasn't really a defined market yet. There was no growth rate to reference. Amazon had to build massive infrastructure for something that most businesses hadn't asked for and couldn't fully evaluate yet. For the first nine years, AWS ran at a loss. Around $5 to 6 billion in cumulative investment before it turned a profit of $1.85 billion in 2015. By 2022, it was generating roughly $22 billion in annual profit.
A BCG analysis in 2009 would have looked at AWS and seen a product with low market share in an uncertain market requiring enormous ongoing investment. It would have flagged it as a Question Mark at best, a Dog at worst. The right call probably would have been to cut it. That call would have been catastrophically wrong.
Netflix in the early 2000s is the mirror image. The DVD-by-mail business was profitable, growing, and dominant in its category. Classic Cash Cow behavior — extract profits, manage carefully, don't rock it. Instead, Netflix cannibalized that business to fund streaming. At the time, broadband penetration was limited and streaming video was unproven. On a BCG grid, moving away from a successful Cash Cow to fund a speculative bet made no sense. What it actually did was reshape the entire entertainment industry.
The Network Effect Problem
The BCG Matrix measures market share as a proxy for competitive strength. For a lot of products, that's reasonable. A brand with more shelf space tends to be harder to displace.
For platform businesses, it misses the actual source of value.
WhatsApp, early in its growth, would have looked like a marginal product in a competitive space. But market share isn't what made WhatsApp hard to compete with. Network effects were. Every new user who joined made the platform more valuable for every existing user. That dynamic compounds in ways that market share numbers don't capture. Once a messaging network reaches critical mass in a given community, the reason to switch isn't price or features, it's who else is on the platform. WhatsApp didn't win by dominating a market. It won by becoming the default.
LinkedIn works the same way. Its value isn't primarily about market share in professional networking. It's about becoming the place where professional identity lives. Recruiters are there because candidates are there. Candidates are there because recruiters are there. That self-reinforcing dynamic is what makes LinkedIn difficult to displace, and none of it shows up in a market share calculation.
The Question the Matrix Doesn't Ask
When you're evaluating a feature or product investment, the BCG Matrix asks: what's your market share? For many platform investments, that's the wrong question.
A better question is: what does this enable?
AWS didn't just sell compute capacity. It enabled thousands of companies to build products they couldn't have built otherwise. That ecosystem of builders created the lock-in, not the market share. A product that becomes infrastructure for other products is in a fundamentally different position than a product that competes head-to-head in a defined category.
Sometimes what looks like a Dog in traditional terms is actually the foundation of a future platform. A low-revenue, low-growth product that happens to sit at the center of user workflows, holds valuable data, or creates switching costs through integration is not the same as a product that's simply losing. The matrix treats them identically.
What to Measure Instead
The BCG Matrix isn't wrong. For traditional product portfolios, hardware lines, consumer packaged goods, it still works. The issue is applying it to businesses where value is generated through ecosystems, engagement, and network dynamics rather than market position alone.
For modern product teams, a few other signals tend to be more revealing:
Retention curves. A product that retains well over time is building something durable. A product where users leave quickly, regardless of acquisition numbers, usually has a fundamental problem that market share can't paper over.
NPS by cohort. Rather than asking whether people like the product today, ask whether users who joined 12 months ago still recommend it. Cohort-level NPS shows whether satisfaction holds as the product matures and as the user base broadens beyond early adopters.
Network effect density. How many meaningful connections does the average user have on the platform? Are users more engaged when they have more connections? If activity scales with network size, that's a sign the core mechanism is working.
Platform activity metrics. For products with ecosystems — APIs, developer tools, integration partners — track activity at the ecosystem level, not just within the core product. Developer adoption, third-party integrations, and API call volume can signal platform health long before revenue reflects it.
The BCG Matrix is a snapshot of competitive position at a single point in time. Most of the interesting strategic questions for platform businesses are about trajectory and dynamics, not position. That's what these signals try to capture.