LEGO didn't almost die from lack of growth. It almost died because of growth — a decade of diversification into theme parks, apparel, video games and lifestyle products that borrowed the brand and diluted the business. The turnaround that saved the company was built on four counterintuitive decisions: divest non-core (Legoland, Clikits, Galidor), cut the unique-brick catalogue nearly in half, turn adult fans from legal adversaries into co-designers, and build a hybrid IP engine where Star Wars pays for Ninjago. Together they restored strategic fit — Porter's term for the reinforcement of activities around a single value proposition — and produced operating margins higher than Apple's at the time. The central lesson: the courage to subtract is the most underrated discipline in strategy.
What was LEGO in 2003?
By the time Knudstorp walked into his first board meeting as incoming CEO, LEGO was a strange organism. On paper, it was still the most admired toy brand on the planet — the plastic interlocking brick invented by Ole Kirk Christiansen in 1932, "LEGO" being short for leg godt, "play well" in Danish. Private. Family-owned by the Kristiansens. Headquartered in Billund, a town that exists primarily because LEGO is there.
But in the decade between 1993 and 2003, under a "diversification CEO" who preceded Knudstorp, LEGO had redefined itself. It was now, according to internal strategy memos, a "children's lifestyle brand" — not a toy company. The implications were everywhere. LEGO owned four theme parks (Billund, Windsor, Günzburg, California). It sold LEGO-branded apparel. It had launched Clikits, a craft line for girls, and Galidor, a range of action figures that broke the brick's sacred interlocking compatibility. It ran a daily LEGO TV show. It licensed a LEGO watch. It was producing LEGO-branded children's books and educational software. It had a video-game joint venture.
The 2003 annual report made the consequence impossible to hide: a loss of 1.8 billion Danish kroner — the worst result in LEGO's 72-year history. Cash was leaving the building at roughly 300 million DKK per month. The Kristiansen family, who owned the company outright, had quietly told advisors they would entertain a bid from Hasbro or Mattel if the bleeding did not stop within 18 months.
The context: how a toy legend nearly went bankrupt
The intellectual move that destroyed LEGO's margin was one sentence: LEGO is about childhood, not bricks. That sentence sounds like brand leadership. It reads beautifully in a pitch deck. And it is the most dangerous thing the company ever told itself — because it widened the competitive set from "plastic construction toys" (in which LEGO had essentially no serious rival) to "everything a child touches between ages four and eleven" (in which LEGO had to compete with Disney, Nickelodeon, Mattel, Hasbro, and every cereal company on Earth).
In strategy terms, LEGO walked away from its strategic fit. Michael Porter's 1996 Harvard Business Review essay "What is Strategy?" — arguably the single most cited strategy paper of the last 30 years — argued that sustainable advantage comes from a system of reinforcing activities built around a single value proposition. Operational effectiveness (being better at the same thing) is necessary but not sufficient; competitors can copy it. What they cannot copy is a tight interlock of activities that all make each other more valuable. LEGO had spent five decades building exactly such an interlock — designer training, molding precision, 50+ years of backward brick compatibility, retail sets that displayed the brick, playroom rituals built around the brick. The diversification decade dismantled that interlock one acquisition at a time.
Parks are a real-estate business. Apparel is a fashion business. Video games are a software business. TV is a media business. Each of them borrowed the LEGO brand while operating under completely different economics, competitive sets, capabilities and talent pools. None of them reinforced the brick. All of them pulled management attention away from the brick. And because the brand name was spread across them, every mediocre Clikits launch or empty Legoland parking lot subtly taxed the equity of the actual product.
This is what Clayton Christensen, in a different context, called the "innovation tax" of portfolio drift — except here it wasn't innovation. It was addition disguised as growth.
The 4 decisions that saved LEGO
Knudstorp was 35, Danish-polite, soft-spoken, and brutally analytical. He had spent the year before his appointment writing an internal memo the family did not want to read: LEGO is not failing because of bad execution. LEGO is failing because it is no longer a brick company. The four decisions below are the operational translation of that thesis.
Subtract — divest everything that isn't the brick
Knudstorp's first move looked suicidal on the revenue line. He announced that LEGO would exit or dispose of essentially every activity that was not the plastic brick. The theme parks went first. In 2005, the four Legoland parks were sold to Merlin Entertainments for roughly US$ 460 million, with LEGO retaining a minority stake and a royalty on ticket sales — a deal that let LEGO keep the emotional halo of the parks without owning the real-estate business. Clikits was wound down. Galidor was discontinued. Apparel was shut down. Video games were out-licensed to TT Games (the studio that would later produce the hugely profitable LEGO Star Wars and LEGO Batman games under a royalty deal). The TV show was quietly ended.
The short-term math was punishing: roughly 40% of revenue evaporated before anyone saw a corresponding drop in losses. The head of marketing resigned. Trade press wrote obituaries. But within eighteen months, cash burn stopped. The family's sovereign-wealth advisors were told to pause the Hasbro conversation. And — this is the part every MBA student under-appreciates — the remaining business became legible again. You could now write on a single index card what LEGO was for. When you can't, your growth will keep destroying your profit.
The principle is Porter's: if a company is bleeding from misaligned diversification, the fix is not smarter diversification. It is subtraction until what remains is the one thing nobody else can do.
Cut the SKU catalogue from ~12,900 to ~7,000
The second decision was, in operational terms, the one that actually saved the margin. Over the diversification decade, LEGO's designers — who were (and are) some of the most talented industrial-product designers in Europe — had quietly expanded the unique-brick catalogue from roughly 6,000 pieces in the early 1990s to about 12,900 by 2003. Every new set tended to introduce a "specialty part" that existed only in that one model. Lightsabers. Specific mini-figure hairpieces. A particular curved slope used once for a pirate ship.
The mathematics of SKU proliferation is brutal, and it was invisible in the P&L because it sat in three different buckets: tooling capex (~€50–100k per new mold), inventory and warehouse cost (~€30k per year of embedded cost for each active SKU), and logistics complexity (every additional part multiplies the risk of the wrong piece ending up in the wrong box, a recurring and expensive failure mode at LEGO). Roughly 35% of unique pieces were being used only once a year.
Knudstorp imposed what internally became known as a "brick budget": no new SKU could be introduced without retiring one. Over five years the catalogue was cut nearly in half, to around 7,000 unique pieces. Inventory days fell from ~102 to around 60. Tooling capex normalized. And, counterintuitively, the sets got better — because designers were forced to use combinatorial creativity (what new thing can you build from existing pieces?) instead of novelty-based creativity (what new piece do we need to tool?).
This is the same lesson Steve Jobs applied at Apple in 1997, when he killed the beige Performa boxes and reduced the product grid to four quadrants: consumer/pro × desktop/laptop. Constraint-based creativity beats novelty-based creativity, almost every time, because constraint forces the designer to reuse assets that the customer has already learned. A child who has 50 basic bricks will build a hundred things. A child who has 500 specialty pieces will build one.
One footnote that matters: Knudstorp's team also ran, between 2006 and 2008, an experiment in outsourcing molding to Flextronics in Czech Republic and Mexico. It was intended to cut unit cost while preserving the bigger catalogue. It was quietly reversed. Brick tolerance at LEGO is ±0.01mm — tighter than most precision-engineering components — and that capability had been built in Billund over five decades. You should not outsource a core capability to save a few cents per kilo of plastic. Most of the molding came back to Denmark and Hungary.
Engage the AFOLs — turn critics into co-designers
The third decision required a cultural U-turn. For years, LEGO's legal department had treated AFOLs (Adult Fans of LEGO) as unauthorized users of the brand: cease-and-desist letters to fan-run websites, refusal to participate in fan conventions, active discouragement of fan-made model archives. The internal view was that adults represented maybe 3% of the market and were a PR headache.
Knudstorp's market research head surfaced a number that broke that view. There were roughly 250,000 active AFOLs globally. They spent about twenty times per person what the average kid spent. They ran hundreds of dedicated blogs and produced more user-generated content about LEGO than any advertising agency could dream of briefing. BrickCon, BrickWorld and a dozen other fan conventions were already happening — whether LEGO participated or not.
Knudstorp launched the LEGO Ambassador Program in 2005 — a formalized relationship with recognized community leaders — and in 2008 launched LEGO Ideas (originally "LEGO Cuusoo" in partnership with a Japanese platform), a site where fans submit designs, the community votes, and sets that cross 10,000 votes enter a commercial review. Fan designers receive a 1% royalty and public credit on the box. The Saturn V rocket. The Apollo Lunar Lander. The 2,079-piece Typewriter. The Bird Project. The International Space Station. All fan-submitted. Several of them are among LEGO's best-selling sets ever shipped.
This is the textbook case of what Henry Chesbrough called open innovation in his 2003 book of the same name: the deliberate decision to let useful ideas flow across the firm's boundary in both directions. In Chesbrough's framework, the closed-innovation firm tries to internalize every step of R&D. The open-innovation firm treats its boundary as porous — licensing in ideas that others develop better, and licensing out or co-creating with communities that can develop what the firm cannot. LEGO Ideas is one of the most-cited real-world examples of the Chesbrough model operating at scale. It also, incidentally, produced the cultural capital that made The LEGO Movie possible in 2014.
The underlying economic insight is simple. When you have a community that spends twenty times more per person than your primary customer, produces marketing content for you at zero cost, and has better taste about what hardcore fans want than your in-house team, suing them is not just ethically strange — it is strategically insane. You turn them into partners, or you pay an agency to try (and fail) to replicate what they already do for free.
Hybrid IP — license selectively, invest in proprietary themes
The final decision was the one that split the board. LEGO's single biggest line entering the crisis was LEGO Star Wars, launched in 1999 under the prior CEO. It was a hit. Now Warner was offering Harry Potter. Marvel was interested. Nintendo was calling about Mario. The purist faction argued that licensed IP diluted LEGO's own universe — City, Castle, Pirates, Technic — and that every licensed sale was partly a payment to someone else's brand. The commercial faction argued that licensed IP was the fastest way to sell boxes to parents who already loved the underlying property.
Knudstorp's answer was neither-and-both. Keep Star Wars. Add Harry Potter (extended 2001). Later add Marvel (2012), DC, Jurassic World, Minecraft, Super Mario, The Lord of the Rings. But simultaneously invest, at equal or greater intensity, in proprietary themes that LEGO would own outright: Ninjago (2011), Friends (2012), Chima, Bionicle (revived).
The economics of the hybrid are elegant. Licensed sets carry a 15–20% royalty bite — painful, but they sell on pre-built global awareness the moment they hit shelves. That cash flow funds the experimentation required to create the next proprietary hit, which carries no royalty at all. Ninjago, when it worked, became a multi-year profit engine that was all LEGO's. By 2015, LEGO's proprietary themes and licensed themes each accounted for roughly half of revenue and a comparable share of profit.
In Aaker's framework this is also an architectural decision: LEGO remains a branded house, where everything carrying the name reinforces the build promise, but within that house LEGO holds a disciplined portfolio of owned sub-brands (Ninjago, Friends, Technic, City) and licensed sub-brands (Star Wars, Harry Potter) that never compete for the master brand's identity. It is a compromise that sounds obvious on paper and is almost never executed well in practice, because it requires saying no to licenses you could afford and yes to proprietary experiments that will often fail. The courage is in the portfolio discipline, not the headline deal.
What LEGO has become
The numbers that came out of the four decisions are, by now, almost folkloric in strategy teaching.
Revenue tripled between 2004 and 2012. Operating margin reached levels higher than Apple's at the time — a fact that is still, for most executives hearing it for the first time, genuinely shocking. By 2015, LEGO had overtaken Mattel as the world's largest toy company by revenue. In 2014, The LEGO Movie grossed US$ 468 million on a US$ 60 million budget — and, remarkably, functioned as a 100-minute advertisement that people happily paid to watch. Knudstorp stepped up to Executive Chairman in 2017. The company remains family-owned, privately held, Danish, and headquartered in Billund — a fact that is itself now part of the brand.
Want to sit in Knudstorp's chair in 2004?
At SMX, you step into the role of LEGO's incoming CEO and make the four turnaround decisions in about 30 minutes. Strategic feedback, comparison with what LEGO actually did, and a counterfactual for every misstep. Free.
Try the LEGO case in the simulator →3 lessons that travel beyond toys
1. Porter's strategic fit — subtract to reinforce
The underappreciated claim in Porter's 1996 essay is not that companies should have a strategy. It is that the strategy must be a system of activities that reinforce each other around a single value proposition, and that the test of whether an activity belongs in the system is whether it makes the other activities more valuable. Legoland did not make LEGO bricks better. Clikits did not make Technic better. Galidor actively made the brick worse by breaking compatibility. Once you adopt Porter's test, divestment stops being a defensive retreat and becomes an offensive move: every non-reinforcing activity you remove makes the remaining activities more valuable. Subtract to reinforce. The lesson applies to any incumbent feeling diversification pressure: before you add a new SBU, ask whether it would make your core stronger — or merely bigger.
2. Brand architecture discipline — branded house vs house of brands
David Aaker's 1991 Managing Brand Equity (and its 2004 sequel Brand Portfolio Strategy) distinguished between a branded house (one master brand stretched across products — FedEx, Virgin, Apple) and a house of brands (independent brands under one parent — Procter & Gamble, Unilever). Both can work. The failure mode is the hybrid by accident: a company that behaves like a house of brands (launching Clikits, Galidor, Legoland, LEGO watches) while labeling everything with the master brand. The master then accumulates the weaknesses of every sub-brand without getting the focus of either architecture. Knudstorp's architectural fix was to collapse LEGO decisively back to a branded house, where everything carrying the LEGO name reinforces the build promise, and to license out or divest the rest. The generalizable insight: a strong brand is not a bigger brand, it is a tighter brand. Every product that wears the name must make the name mean more, not less.
3. Open innovation as operating system
The third lesson is Chesbrough's. For most of the 20th century, companies treated their firm boundary as a wall: inside was R&D, branding, manufacturing; outside was competition. Chesbrough's argument, articulated in Open Innovation (2003), is that in a world of distributed knowledge and abundant passionate hobbyists, the wall becomes a liability. The companies that win are the ones that build porous boundaries — platforms where external communities contribute meaningfully to the firm's innovation pipeline in exchange for recognition, royalty, or influence. LEGO Ideas is a canonical case. So is Threadless in apparel. So, in a different register, is GitHub in software. The strategic implication: if there is a passionate community that knows your product better than your best employees do, your job is not to control them; it is to build the platform on which they can contribute without the firm losing its coherence. Community is not a marketing channel. Community is a factor of production.
Conclusion: why this case still matters
It is tempting to read the LEGO turnaround as a story about toys, or about a uniquely admirable family firm, or about a brilliant CEO. It is a better story if you read it as a story about the most underrated discipline in strategy: the courage to subtract. Almost every framework an MBA student is handed — growth matrices, Ansoff matrices, adjacency maps, blue-ocean canvases — is implicitly biased toward addition. More products. More segments. More geographies. More channels. The assumption is that growth is always the right answer and the only question is where.
What LEGO teaches, and what Porter, Aaker and Chesbrough collectively teach, is that growth is sometimes the disease. When a company's activities stop reinforcing each other, growth amplifies the misalignment. You don't fix it with smarter diversification. You fix it by subtracting until what remains is the one thing nobody else can do — and then you reinforce that, relentlessly, across operations, brand, product, community and IP.
You don't learn this by reading. You learn it by making the decision.