You Should be a Cannibal! | The Discipline of Innovation

You Should be a Cannibal!

I was doing some work with a company this week, and ran across one of my pet peeves. The organisation itself was very exciting. We interviewed over 20 people about various projects that were going. To a person they were smart and engaging, with great ideas, vision and energy. It was invigorating. The number of exciting projects that they have on the go is enormous. So what’s the problem? The problem is that many of these projects will never get off the ground. Why? Because they might cannibalise existing business.

This is the worst possible excuse for killing innovation. In many ways, if you worry about cannibalisation, you are falling into exactly the same trap that makes Net Present Value project evaluation so flawed. Earlier this week, Ralph Ohr and Wim Rampen both pointed me to an article that illustrates this flaw perfectly. The article is called Innovation Killers, by Clayton Christensen, Stephen Kaufman and Willy Shih from the January 2008 edition of Harvard Business Review. They include this diagram:

This shows the problem perfectly. NPV analysis assumes that our current level of performance will remain stable and that is the probable future that is used to evaluate new innovative projects. However, the ‘do nothing’ option does not result in staying at the same level of profit and performance. If we do nothing, our performance will decline. We will be worse off because someone out there is figuring out how to destroy our market. If we try new things, and they work, we might stay ahead of them. If we don’t, they’ll knock us over. There’s no way around it.

This same flawed assumption is being made when people worry about introducing new ideas that will cannibalise their current product or service lines. Of course they will. But again, the do nothing option does not result in ‘everything stays the same’. The do nothing option results in someone else doing the cannibalising.

Halfway through the day, our host said this:

When we compare ourselves to our current competitors, we look pretty good. We’re innovative and ahead of the game. But that’s the wrong comparison. I’m not worried about the people already in our market. I’m worried about the people I’ve never heard of.


The simple fact of the matter is this: we must introduce innovations that take away our current market share. There’s no such thing as a cash cow – if you have a product that is dominating the market, but you’re not innovating around it, it’s not a cash cow, it’s already dead.

Cannibalise your current products. Kill your best performers, or someone else will kill them for you.

About Tim Kastelle

Student and teacher of innovation - University of Queensland Business School - links to academic papers, twitter, and so on can be found here.

13 Responses to You Should be a Cannibal!

  1. Sam 14 March 2010 at 4:28 pm #

    Nice framework!

  2. Tim 14 March 2010 at 5:23 pm #

    It is, isn’t it? Half of Christensen’s stuff is fantastic, and the other half is…… less fantastic. It’s a bit weird.

  3. Sam 15 March 2010 at 8:10 am #

    Yeah true. One thing that struck me was that the framework didn’t seem to incorporate any notion of innovation blowback (whether it’s dollars, reputation etc))? I know the model is meant to be illustrative but…the assumed up swing in cash flow from introducing a new innovation seems way too positive. Assuming away blowback could account for this and also explain why many of us innately risk averse humans aren’t constantly innovating. I suppose it also highlights why it’s crucial to get a high quality “fail fast” process in place.

  4. Tim 15 March 2010 at 11:59 am #

    Well, it is and it isn’t too positive Sam. Any time you evaluate a potential new scheme, if the potential upside isn’t positive, you don’t do it. So one of the assumptions here is that everything being evaluated has at least gotten over that hurdle.

    The next question is how accurate are the projections? In many cases this is just pulling numbers out of thin air. You do want to try to make them as accurate as possible, but they’ll always be fuzzy. Especially for more disruptive innovations…

  5. Andrew Dick 15 March 2010 at 8:48 pm #

    Do you think part of the reason young businesses are more innovative is because they have less “sophisticated” finance functions and therefore don’t get into detailed, innovation killing DCF analysis?

  6. Tim 15 March 2010 at 9:02 pm #

    Quite possibly Andrew. Although the other reason is that they often have less choice – they have one idea and they have to get it to work or they’re out of business. In that circumstance, it’s easier to ignore financial forecasts, especially if they’re not so favourable…


  1. uberVU - social comments - 15 March 2010

    Social comments and analytics for this post…

    This post was mentioned on Twitter by _innovation: RT @timkastelle – You Should be a Cannibal- kill your best products before someone else does #innovation…

  2. Innovation Lessons from Charles Leadbeater « Innovation « Innovation Leadership Network - 15 March 2010

    [...] is a good discussion of why the organisational imperatives within large firms lead to incremental innovation rather than breakthroughs. It is much easier to get sign off on an [...]

  3. What Does a Good Innovation Option Look Like? « Innovation « Innovation Leadership Network - 17 March 2010

    [...] essay on the NPV trap, which is very supportive of the survey results. Tim subsequently wrote a blog piece on this [...]

  4. The Perils of Disruption « Innovation Leadership Network - 21 May 2010

    [...] Here’s how I put it in an earlier post: [...]

  5. The Perils of Disruption « gabriel catalano | in-perfección - 21 May 2010

    [...] Here’s how I put it in an earlier post: [...]

  6. How Do You Know When to Jump? « Innovation Leadership Network - 25 October 2010

    [...] There can be good reasons to not react to a threat. However, whenever I hear someone say that they won’t introduce an innovation because it will cannabilize market share, I get worried. This is another “internal constituency” problem – over the long run, you should be a cannibal. [...]

  7. Blogging Innovation » Better to be a Cannibal - 1 November 2010

    [...] There can be good reasons to not react to a threat. However, whenever I hear someone say that they won’t introduce an innovation because it will cannibalize market share, I get worried. This is another “internal constituency” problem – over the long run, you should be a cannibal. [...]

Leave a Reply