7 Powers

Intel ran two businesses with the same founders, plants and balance sheet, and was first into both. Memory chips ended at nothing and microprocessors at about $150B — in Helmer’s words, “$0 for memories and $150B for microprocessors”. So whatever split them was nothing the two shared: not talent, not timing, not money.

Helmer’s answer is that competition wore one down and not the other. He is an economist and expects the wearing: rivals copy whatever earns more than its cost of capital until it no longer does. IBM, Motorola, AMD and a long list of others spent billions on microprocessors and still could not do it to Intel. Whatever stops the copying he calls Power: “the set of conditions creating the potential for persistent differential returns”.

Every Power has two halves, and only one is rare. The Benefit lifts cash flow through a higher price, a lower cost or less capital; every cost-cutting drive has one. The Barrier stops rivals copying the Benefit away, and almost nothing has one. Hence his first rule: “Always look to the Barrier first.”

A Barrier has to hold against each rival on its own. Helmer calls Power “an explicitly relative concept”: a firm can hold it against one competitor and not another, and the list includes rivals that do not exist yet and rivals meeting the same need with a different product. One rival that can copy is enough to wear the margin down.

The Barrier matters more than the size of the Benefit because most of a firm’s value lies in the far years. Helmer writes value as market size, times discounted growth, times long-run share, times long-run margin above the cost of capital. On a common valuation model, he notes, for a company growing 10% a year the next three years account for only about 15% of its value. A good spell that fades is worth little.

From that sum he gets a one-line definition of a company’s strategy, which he calls the Mantra: “a route to continuing Power in significant markets”. Market size and growth are the significant markets; share and margin are the Power. He claims the seven forms of Power in the next seven chapters are the only ones there are: they covered every one of the more than 200 strategy cases he led.

Scale becomes Power when a large cost does not grow with sales, and Netflix made its biggest cost behave that way on purpose. Once it paid for exclusive and original shows instead of renting them by use, every streamer had to pay the same sum whatever its size. Helmer’s sum: if House of Cards cost $100M and Netflix had 30M subscribers, it cost “three dollars and change” a head. A rival with one million subscribers would pay $100 a head for the same show.

Fixed costs are the commonest source but not the only one. Helmer lists four more: tanks and warehouses, whose cost tracks their surface while what they hold tracks their volume; delivery routes that get cheaper as they fill, the problem facing any new rival to UPS; learning that grows with output; and the buying power that helped Wal-Mart. In each case the Benefit is the same, a lower cost per unit.

The Barrier is subtler. The smaller rival can close the cost gap only by winning share, and winning share means a better price. The leader sees the move and matches it, and from its lower cost it can afford to. After a few rounds the follower learns that chasing share loses money, and stops. Intel against AMD is his long-running case: a great business for one and “persistent pain for AMD”.

To size the lead, Helmer asks what margin the leader earns when prices fall to the point where the follower earns nothing. He calls it the Surplus Leader Margin. For a fixed cost it is the fixed cost as a share of the leader’s sales, times the leader’s sales over the follower’s minus one. The first term describes the industry, the second the firm’s place in it. If either is zero there is no Power, however heavy the fixed cost.

That split, industry economics against competitive position, runs through every chapter that follows, and Netflix worked both. Originals changed the industry’s cost structure; an early, careful start gave Netflix the subscriber lead. Had it taken the industry’s economics as given, Helmer argues, streaming would have offered it no route to Power at all.

A network effect protects only the network that users care about. In 2010 BranchOut set out to beat LinkedIn, then at 70M members, by building on Facebook, whose user base was almost ten times larger. It raised $49 million, grew to about 14 million users by the spring of 2012, and then collapsed. Hearst bought its assets and team in September 2014.

Facebook’s size did BranchOut no good because users kept work and private life apart. Recruiters go where the most professionals are listed, and professionals list where the recruiters go; in that network LinkedIn led and Facebook barely counted. Helmer found “very little spillover” from one to the other. Facebook learned the same lesson when Facebook at Work failed.

Where the effect is real, it is close to unbeatable. The leader’s product is worth more because more people use it, so it can charge more, and a follower must cut its price so far to close the gap that the price can go below zero. Helmer guesses BranchOut would have had to pay each user to leave LinkedIn. Such markets tip: once one firm leads, the others give up, and even Google could not unseat Facebook with Google+.

The catch is that a network effect is not Power until it is big enough to pay for itself. “This is the problem I see often in Silicon Valley,” Helmer writes: firms assume a network effect, spend heavily to reach the tipping point first, and only then learn how large the effect is and whether it can earn money. His example of a company that met that problem is Twitter.

Helmer’s own idea, and his favourite, is a Power that uses the incumbent’s success as the barrier. His case is Vanguard. John Bogle won his board’s backing on 1 May 1975 for a fund that would simply track the market. It launched in August 1976 with $11M and held only $17M by mid-1977. More than a decade passed before it took off; by the end of 2015 Vanguard managed more than $3T.

He defines it as a newcomer adopting “a new, superior business model which the incumbent does not mimic due to anticipated damage to their existing business”. Vanguard dropped the stock pickers, cut sales and trading costs, and passed the saving to savers as higher net returns. Fidelity had every skill needed to copy it. What stopped Fidelity was its own sums: cheap index funds would have eaten its high-fee active funds. “The Barrier, simply put, is collateral damage.”

A slow incumbent is not always a counter-positioned one. Kodak saw digital photography coming and “spent lavishly to explore survival options”. But digital film was semiconductor storage, where Kodak brought nothing, so the new business held no Power for Kodak whatever happened to film. Helmer keeps that case out of the category: digital was a poor bet for Kodak on its own, damage or no damage.

The damage that deters an incumbent comes in three kinds, which he names “Milk, History’s Slave and Job Security”. Milk: the new model would cost the old business more than it brings in, so the incumbent harvests what it has. History’s Slave: past success colours the view, as when Fidelity’s Ned Johnson asked, “Why would anyone settle for average returns?” Job Security: the people deciding are paid on the business a response would hurt. Often all three act at once.

The odd result is that the higher the incumbent’s margins, the stronger the challenger, since high margins mean more to lose by cutting into them. Helmer advises the challenger to speak of the incumbent with respect instead of boasting, because a slower read on the threat buys time. The Barrier wears away all the same. As the old business shrinks and the new one proves itself, the damage from answering it falls, and the incumbent at last moves.

It is a partial Power. It protects a challenger only from the incumbent, not from other challengers with the same model: In-N-Out holds it against McDonald’s but not against Five Guys. Nor is it exclusive, since several challengers can be counter-positioned against one incumbent at once. A challenger still needs some other Power against its own kind.

Helmer sets out, half in jest, the stages an incumbent goes through: denial, ridicule, fear, anger, and “Capitulation (frequently too late)”. Nokia led smartphones in 2008 and had left the market by 2014. He puts its chief executive, Stephen Elop, at the anger stage, quoting his 2011 memo: “The first iPhone shipped in 2007, and we still don’t have a product that is close to their experience.”

Customers will keep paying for software they dislike if leaving costs more. SAP is the case. In one study, 43% of 588 SAP customers were unhappy with its response times; in another survey of more than 1,000 customers, 89% expected to keep paying the annual maintenance fee.

HP shows what leaving costs. In 2004 it moved its North American server sales, then $7.5 billion of revenue, onto SAP. The executive in charge had run five such moves before; she allowed three weeks for the changeover, held three weeks of extra stock and booked spare factory space. Even so, as many as 20% of server orders stuck between the old system and the new, and HP’s chief executive put the cost at $160M. Switching cost far more than the software, and nobody could have said in advance how much.

Helmer defines Switching Costs as the value a customer expects to lose by moving to another supplier for further purchases. They come in three kinds: financial (new licences and the software around them), procedural (retraining, and the risk of errors in moving data) and relational (ties to the vendor’s people and to other users). SAP deepens all three by adding products and buying companies.

The Benefit is narrower than it looks. It applies only to follow-on sales to customers a firm already has: there is none on customers not yet won, and none if there is nothing more to sell. Rivals can count too, so in a mature market they bid the worth of a locked-in customer away at the point of sale. The money lies in winning customers before that bidding starts, which Chapter 9 places in takeoff.

Nor does it belong to one firm. IBM and Oracle keep their customers just as SAP does, and position is all or nothing: a firm either has a given customer or does not. A shift in technology can also end them, which is why, Helmer notes, SAP and Oracle are working not to be leapfrogged by cloud software.

Sometimes the Power is a group of people. Pixar’s first ten films averaged 94% on Rotten Tomatoes, with only Cars below 90%. Toy Story cost $30M and took over $350M worldwide. Steve Jobs bought the company from George Lucas for $5M in 1986; Disney paid $7.4B for it in 2006. No director or film series had matched that run.

Helmer traces it to the Brain Trust, the team whose bonds formed in the hard making of Toy Story and the two films after it. He defines the Power as “Preferential access at attractive terms to a coveted asset that can independently enhance value.” The Barrier is what he calls “fiat”: a patent, a property right or, at Pixar, personal choice. When Disney tried to hire John Lasseter back in 1988, he said: “I can go to Disney and be a director, or I can stay here and make history.”

Because so many things look like scarce assets, he sets five tests. Idiosyncratic: if a firm keeps getting good assets cheaply, the real resource is whatever lets it. Non-arbitraged: Brad Pitt lifts a film’s takings, but his fee takes most of the gain. Transferable: the asset should create value elsewhere too, as Pixar’s leaders did when they revived Disney Animation.

Ongoing: the resource must still be producing the advantage. Spencer Silver’s weak glue made Post-it notes possible, but once the product was established 3M’s returns rested on its patents and trademark. Sufficient: it must explain the result on its own. George Fisher ran Motorola well and could not rescue Kodak, so he was not the resource. At Pixar, new directors brought into the group were often replaced — on Toy Story 2, Ratatouille and Brave — which points to the group, not any one member.

Helmer saves the rarest Power for last. In 1969 Toyota held about 0.1% of the US car market against General Motors’ 48.5%. By 2014 it had drawn nearly level with GM and Ford, and GM had fallen below 20%. The cause was a production system Toyota had been building since 1950, when Eiji Toyoda spent three months at Ford’s River Rouge plant, judged its deep stocks wasteful, and preferred the supermarkets that restocked only when a shelf emptied.

GM had every chance to copy it. In 1984 it opened NUMMI with Toyota in Fremont, sent the workers to Japan to train, and soon had defect rates close to Toyota’s own. Toyota hid nothing. Yet GM could not repeat the results in its other plants. A GM plant manager said: “the one thing you don’t see is the system that supports the NUMMI plant.”

The Barrier is the time it takes, for two reasons. Complexity: even the Model T took 7,882 assembly steps, before counting the supply chain behind it. Opacity: the system grew from the bottom up, much of it was never written down, and it took Toyota itself fifteen years to pass it to its own suppliers.

So Helmer agrees with Michael Porter that operational excellence is not strategy: what can be copied earns nothing lasting. His formula is “Process Power equals operational excellence, plus hysteresis”. The experience curve does not undo this. If costs fell with experience across firms at any one moment, a firm twice the size of its largest rival would hold an operating-margin edge of 15% to 30%. Gaps that large are rare, which shows how fast rivals copy.