Good To Great by Jim Collins

I recently finished reading “Good To Great”, which is a bestseller on how good companies are different from great companies in terms of their management structure.

First, a quick review – after reading this book, I can’t help but get the sense that someone has baked a cake and covered it with turd. This may be symptomatic of all management bestsellers – it contains a few good insights covered with a lot of made-up terminology and rhetoric. I’ll try to pull out the key ideas in the book – brushing off the frosting, as it were.

Level 5 leadership

Two types of CEOs are compared – superstars like Iacocca aren’t all that hot, as the companies couldn’t endure. The real superstars are the CEOs you don’t hear about. They have the following qualities:

  • Humble
  • Takes the blame, but externalises to luck or others for credit
  • Solid resolve. This is necessary for two reasons – getting results takes years, and you don’t want anyone to think you’re a wuss.
  • Ambitious for the institution, rather than themselves. (This is the opposite of what the military might call a “careerist”.)
  • Accepts the truth rather than self-congratulation, but has faith that you will prevail in the end.
  • Is very disciplined in getting to their goal, cutting at details to get the answers.

Having a genius up top is great, but they will be surrounded by yes men who are just executors of will. Once the genius is gone, they don’t know what to do and the company fails.

An odd contradiction is that people who are ambitious and shout the loudest are most likely the ones that will make it to the top of the heap. So we need to watch out for these “Level 5” leaders.


The importance of getting the right people is highlighted over and over in the book. The argument goes like this – if you have the right people, you don’t need to bother doing a whole bunch of stuff:

  • Motivating people
  • Aligning people
  • Playing with your incentive structure
  • Making people passionate
  • Enforcing discipline

All you need to do is make sure you don’t stamp over these factors, e.g. don’t demotivate motivated people.

When recruiting people, here are the rules:

  • You should look for character and work ethic, and not worry about skills and experience – ethic can’t be taught easily, but skills can.
  • Don’t hire just because you need to, just keep looking
  • Boot the wrong people
  • Give the best people the biggest opportunities to work on

In a company where this is happening, people will either stay a very short time or a very long time.

What the leadership team should do

  • Ask a bunch of questions
  • Debate (and don’t coerce anyone)
  • The leader decides what to do (the group doesn’t have to come to consensus)
  • Analyse your results and autopsy failures


  • Find one simple concept and stick to it. (This is similar to the principle of Curly’s “one thing” from City Slickers.)
  • Cut out anything that isn’t working and / or doesn’t fit with your central concept.
  • Don’t change your strategy in response to a competitor.
  • Do change the manifestation of your strategy if necessary – i.e. you can change your product lines, but don’t change what you’re trying to deliver to the customer.
  • Technology accelerates things, but it’s not the basis for a strategy.

Use these ideas to pick your concept:

  • What can you be the best in the world at?
  • What drives your economic engine?
  • What are you passionate about?

Two quick notes – once you understand your economic engine, you can play around with your metrics. Are you looking for revenue? Returns on assets? Growth? Pick one and roll. Delivering shareholder return is not allowed, as it should be a side effect of whatever you’re doing.

Secondly, the question of what you’re passionate about depends on the people you already have, as it’s tough to get people to be passionate about whatever.

The end!

Beyond Collaborative Marketing

There’s a ridiculous amount of content to read on the rise of collaborative marketing – the idea that instead of a mass media source talking to the masses, internet media will open a dialogue with everyone. This is just asking marketing to wake up and realise what’s already happened – people see your ads, but then go online and read forums and blogs before purchase.

This is how it is

People go online to get a sort of word of mouth marketing. They might look up product review sites, shopping comparison websites, or even check out the company on Wikipedia. But bear in mind that there are two categories of people who go online.

  • People who post content
  • People who read content

So, we can see an evolution of how information gets from place to place.


So we can see how suddenly, the folk in the middle – what a marketing person might call “key influencers” – have become incredibly important.

Make your point or go home

So here it is. The big question is – how does the company influence the people in the middle, and to do what?

I have diddly squat data to back this up, but here are my assumptions.

  1. Only certain types of people are content creators / key influencers.
  2. As they become more important, companies need to raise the rewards for those people to interact (and positively) about the company’s product / service.

Repeat after me

For people who just need to talk – reviewers, bloggers and so on – that can create a conflict of interest. On one hand the company can offer increased access to products and people, but if that results in negative reviews, that’s a problem.

I consider geek marketing a glimpse into the future of what might happen. For a long time, I built and configured my on desktop hardware, and I was pretty serious about buying the best hardware to put in my box. I’d go to hardware review sites like Tom’s Hardware to do so.

Slashdotter Savannahlion used to do the same thing, but then he switched.

“As near as I can figure, Tom’s Hardware reviews read too much like brochures. It’s just enough to try and get a person interested. But whatever it is, it’s enough to prevent me from utilizing his site for anything more than keeping track of the latest hardware. Nowadays, if I want real hard and honest opinions, I just hop on over to my favorite forum/BBS/IRC/whatever and sort through the flamefests to get a feel for a particular piece of hardwares viability.”

So the top – the company – exerted too much influence. The chain here is slightly more complicated:


Here’s the critical difference – what does the consumer perceive as “us” and “them”? Who do we trust? The line used to be just below the company – suddenly, it moved down to cut off the larger influencers.

Rolling Your Own

Companies don’t just interact with influencers in order to get the word out. They can also get them to help create the product (hello Open Source).

One interesting example my Tech Marketing professor, Mo Sawnhey, gave us in class was this one. A drinks company asked their customers to help create a flavour for their product, and then offers to pay them royalties on their product. It’s an interesting example of CPG crowdsourcing with high reward for participation – and increasingly, high rewards will be necessary to get this limited pool of people to participate.

The danger is that as the rewards go up, instead of customers who are content creators, we’ll just get content creators – a sort of middle group of freelancers. They key is that in this case, they’re not customers any more, and may have some different insights than the customer base.

No Fate But What We Make

So – what can we conclude?

  • The old days are over. Understand that most of your shouting power is now outside the company.
  • Influencing skills are getting more important, so having some would be awesome.
  • Don’t go too far – you may neuter the influence of the key influencers, like a cat.
  • If you get negative noise from your influencers, don’t withdraw your support – just fix the problem.

Admiral Tom Zelibor at Kellogg

I had the pleasure of meeting a Naval Rear Admiral today (specifically this one), who’s now a Dean at a Naval War College. I resisted the temptation to call him “sir”, and we chatted about his views on leadership. Here‘s Kellogg’s note on the same.

He was wonderfully direct in his communication. Here are some of the insights I picked up.

  • Coaching is the most important part of a senior leader’s job. One should be training those who report to you ultimately to replace you.
  • Leaders should bear in mind what’s best for the organisation, not what’s best for oneself. The two can often be in conflict.
  • He shared a great quote from a Lieutenant General, who said that a leader should be prepared to take the heat for the mistakes of subordinates. That’s awesomely direct – most people really don’t want to do this, but I can see why it’s necessary.
  • Subordinates who are given real responsibility will work their arses off for you.
  • Self-reflection is necessary. This struck me as a most un-military thing to do, so it’s especially meaningful that he said it.
  • One must adapt one’s communication style to others in order to influence them.
  • Don’t have too many checks and controls for decision making. This one is tough for me – as an engineer, I like process, checks and controls. According to Tom, that equals policy which equals bureaucracy. My counterpoint is that it helps minimise the effect of some crazy leader making huge gambles that are out of line with the feel of the whole organisation. Tom believes it slows down decision making. I think we’re both right – the question is where a balance can be found. I suspect that if you trust your people, only minimal checks are required, if at all.

Over the course of our meeting, Tom described a little of his time as a Commander Air Group, or CAG. As he explained what a CAG was, I held myself back from yelling “ooh! I’ve seen those on Battlestar Galactica!” Immediately it become apparent how far apart the worlds of technology and the military are. Perhaps that makes sharing our learning even more useful.

Beverly Hills Cop and the Diffusion Framework

One of the most useful frameworks I’ve learned at Kellogg is Geoffrey Moore’s diffusion framework. One of my favourite people is Beverly Hill Cop’s Axel Foley. I was delighted to discover that I could combine the two.

Most people in tech have, at some point, flicked through “Crossing the Chasm”. It describes how a new technology can diffuse from a neat technology demo into something people use every day. There are five sets of people involved in making that happen. The first segment persuades the second, the second the third, and so on … and it all starts with an idea. Axel’s idea.

Axel is an innovator.
You have a very big mouth
“You have a very big mouth, sir.”

Axel doesn’t get things done in the usual way. He comes up with new ways of solving problems (crimes). He has a pretty interesting theory about who killed his friend Michael Tandino, but he needs to convince a whole series of police officers in order to secure a conviction for the killer. The first one to believe him is Billy.

Billy is an Early Adopter.
Running out of ammo
“You know what I keep thinking about?”

Billy is the first of the Beverly Hills Police Department to believe Axel’s theory about Victor Maitland being a drug dealer and smuggler. He doesn’t come up with great ideas himself, but he can see the potential in great ideas when they arrive, and is adept at understanding their implications.

(Usually, an Early Adopter is organised and mercenary about getting deadlines and project plans in place. In this case, Billy is a bit of a scatterbrain.) Next in line is the Captain.

Captain Bogamil is part of the Early Majority.
Is this what really happened?
“Is this what really happened?”

This transition is critical. This is what Moore refers to as the “chasm”, because it’s so hard to get Bogamil to accept Axel’s theory. There are two reasons why – Bogamil likes doing things by the book, and sees no reason to change. If there’s cause for a conviction, it should be presented in the same way as usual. Bogamil wants “evolution, not revolution”. Worse – Bogamil’s buddies are all other cops like Bogamil. He doesn’t really listen to what Billy has to say.

Forget what you can prove ... talk to me.
“Forget what you can prove … talk to me.”

So how can Bogamil be convinced? In this scene, Axel suddenly drops his act and starts talking like a police officer. Suddenly, Bogamil can relate to his theory and asks Taggart to look into the problem, not allowing a Detroit detective to solve the crime. His solution is inherently practical, and that’s the key to selling to the Early Majority. They want a practical solution to the problem. At the same time, Bogamil can see the value in Axel’s theory.

Taggart is part of the Late Majority.
You're going to pay for this.
“You’re going to pay for this.”

Taggart is not big on theories. He’s practical, sure – but remains a sceptic even at the point where they are invading Maitland’s house and shooting bad guys. He would prefer it if Axel had never arrived in Beverly Hills – but by the last scene, where’s it’s clear that Axel’s theory is the truth, he has accepted the state of affairs. The last scene brings back the Chief.

The Chief is a Laggard.
You expect me to believe that report?
“You expect me to believe that report?”

The Chief is staunchly opposed to Axel and his theories, not even stopping to listen to what he has to say. He will pay a price (ignorance) to avoid Axel’s theory, and in the end has to be lied to in order to avoid acceptance.

From each stage, the previous group is the strongest influencer of the next. Billy knows he will never be able to convince Taggart of anything. At the end, the Chief asks Taggart to tell him what actually happened – as a Laggard, his closest reference is a Late Majority user like Taggart.

I hope this illustrates how the Technology Diffusion framework can be used to write classic 80s films. All we need now is Harold Faltermeyer.

De Beers Case Study

Carrying on the MBA content series, I’d like to highlight a case study we ran through last term on DeBeers. It’s a pretty standard study that I expect a lot of business schools use. I hope you’ll find it illustrative.

Almost every business school student knows that the value of diamonds is vastly inflated, and in large part that’s due to the work of the DeBeers company. In a nutshell, here’s how they did it:

  • Through a series of nifty deals, they chiefed all of the diamond mines they could lay their hands on. Initially, these were all located in the same area, making this possible. This gave DeBeers a virtual monopoly on diamond production.
  • DeBeers mined at a regular rate, but stockpiled most of it, only letting a trickle out into the market. This artificial scarcity is what drove prices up.
  • If another mine came online that didn’t play ball, DeBeers would flood the market with specifically the type of diamond that mine produced, driving down their revenue stream.
  • Then, an awesome marketing campaign. The phrase “Diamonds are Forever” is a product of the DeBeers company. As a result, engagement and wedding rings almost always have one or more diamonds in.
  • They then moved to control the distribution chain, by being pretty crappy to jewellers. Since DB had monopoly of the product and created demand, the distribution chain were left with little leverage.

So, everything was hunky-dory, until a few things happened:

  • Those pesky miners kept discovering mines in different places. DeBeers normally just bought the things, but they couldn’t keep on doing that.
  • As soon as their production monopoly started to disappear, distributors got an incentive to go elsewhere.
  • Since the marketing campaign was about diamonds rather than DeBeers, suddenly their traditional marketing campaign was working for their competitors.
  • Everybody went to see Blood Diamond.

The answer?

  • DeBeers was and is awesome at marketing, so they shifted to market DeBeers diamonds instead of diamonds generally.
  • They certified that their diamonds as conflict free.
  • And the biggie – they opened retail stores, and started making jewellery.

And I found myself walking past the DeBeers shop in London the other day. Here’s what it looks like:


Was going into retail the right direction? There’s no easy answer – they are now able market directly to the consumer a product that can be tied directly to a store and a retail experience, but they are in direct conflict with their own distribution channel.

Time will tell!

Scope Extension and Acquisitions

Carrying on the MBA content series, I’m going to talk a little about why a company might want to acquire new capabilities, either by buying companies or creating a competency internally.

So, let’s take the example of a vertical chain (i.e. a chain inside one industry):

Supplier -> Company -> Customer

For example, this could be:

Textile supplier -> Clothing manufacturer -> Retail chain

From here, there’s two things the manufacturer can do. It can buy its supplier (an action called “backward integration”), or it can buy its customer (an action called “forward integration”). Or both, if it’s feeling flush. We’ll look at backwards integration in this example.

Why might this be a good idea?

The first is simple and obvious – you can stop supplying any of your competitors with textiles. A note of caution here – you can usually negotiate an agreement that does the same thing, which is a hell of a lot cheaper.

The second is that it can optimise the chain. It can start to produce just the right sort of textiles – maybe they’re pre-cut into shapes – to fit really well into its manufacturing process. It knows that by getting optimised product, it can streamline its internal operations and save money.

“But hold”, I hear you say. “Why does it need to buy a supplier to do that? Can’t it just ask a supplier to do it? It has a good relationship, right?”

“You’re quite right,” I say. “But there’s a complication.”

The hold-up problem

Let’s say that the textile supplier serves a few different clothing manufacturers. It receives this request for customisation, and it wants to maintain a good relationship, so they think about it. But one manager with his head screwed on straight says, “Hang about – if we customise our product for you, you’ll be able to negotiate really good prices with us, because you know we are depending on you for our business. So I’ve thought about your request, but you can bugger off.”

The reason for this bad language is that there is a difference between market price and price to a specific customer. If the difference is huge, this creates what’s called a hold-up problem. And suddenly, the supply -> manufacture part of the chain is sub-optimal. This is something that extending scope can fix.

As always, there is one exception, which is people (or “human capital” as they call them round here). Supposing there’s one guy who makes cool patterns you put on your clothing in the manufacture process. You can ask him to come and work for you, and work in a specific way that optimises your internal operations. If this guy has his head screwed on straight, he’ll say, “Hang about – if I change the way I work, you’ll be able to negotiate really rubbish pay deals with me, because you know I can now only work at your company. So I’ve thought about your request, but you can bugger off.”

And that’s fair enough, because you can never nail down people in the same way you nail down machines. They will stick it to the man.

Employee Turnover

Please welcome a new post category, MBA Content.

When searching for an MBA, I often wondered what on earth people do for two years, and this category will help answer that question. I’ll be including a few things that we’ve been squirreling away on that I’ve found particularly interesting or surprising.

This first note definitely fits into the surprising category, and it’s about employee retention. (Not anal retention.) The “mom and apple pie” line on retention says that you should love your employees, make them feel cared for, and you are doing well if they love you back. See Apple. See SAS.

So – why has it worked for them? Being nice to employees means that they stay. That’s an advantage in an area like software – you only have to take a quick look at the Daily WTF to figure out why. When people leave, they take a lot of knowledge with them, and it takes a long time to train someone new up and be as productive as someone who is experienced.

Someone who has been working for a software company for a long time is worth MORE than someone who hasn't.

Now let’s look at McDonalds. The employee churn rate is, obviously, very high. Now we can deduce why the Golden Arches don’t want to spend money on extra benefits for their burger flippers – it’s incredibly easy to train a replacement. (Remember Lester Burnham.)

Someone who has been working for McDonalds (on the front line) for a long time is worth NO MORE than someone who hasn't.

So, this is where “mom and apple pie” doesn’t work. A company like McD (or Foxtons, or whoever) should be putting money into anything other than benefits, teaspoon races and group hugs for their employees, whereas any software company worth their salt should be providing free transport, gourmet food and beanbags at every corner.

Google doesn’t seem so fuzzy now, does it?