Astro Teller works at Alphabet's X where he is director of innovation (or some such title).


Each participant is assigned two "innovation formulas" which you should be prepared to talk about in the next class.

Assignment. Meet Astro Teller. Watch "How to Drive Innovation Inside Organizations" video and wait for your assigned mantra/formula. Write a brief explanation of what it means. Identify one example of the negative side of the equation in higher education as we know it. Cook up a positive side alternative. You may also want to view Teller, Astro. 2016. Systematize Innovation or "Astro Teller Interview - Astro Teller's Top 10 Rules For Success"1

  1. $Innovation(Someone) \gg Innovation (everyone) \gg Innovation (no one)$ (4:30)
  2. $!Story \implies !Innovation$ (8:10) [DD, AME]
  3. $\text{Innovation} (story) = \text{Benefit} (vision) + \text{Product} (vision) + \text{How} (vision) + \text{WhyNow} (vision)$ [MAR, TJS]
  4. $100 \times Innovation(1) \gg 1 \times Innovation(100)$ (15:30) [AMB, MC]
  5. $Rebuild = Innovate$ [AME, ELH, SLS] (27:22, 47:45)
  6. $\text{Technology Judo} \gg \text{Technology Kung Fu}$ [CAHC, SK] (29:20, 1:10:00)
  7. $\Delta(Assumptions) \implies Innovation$ (36:30) [SK, NMN]
  8. $!Understand Details + !Delegate \implies !Innovation$ [HAC, CCC, ANS]
  9. Make the technology problem go away. Technology is not the hard part. (41:00)
  10. $Effort(Innovation_{V2.0}) < Effort(Innovation_{V1.2})$ (46:30) [ELH, CAHC]
  11. $P(Innovation \mid Bob) > P(Innovation \mid Idea_Z)$ [SCP, AKRD] (1:07:00)
  12. $\text{Idea Transplant} \implies Innovation$ [ANS, SLS]
  13. $\text{New Tools} + \text{New Metrics} \implies Innovation$ [TS, AMB, TJS]
  14. $CleanUp(\text{painful workarounds} + \text{rule breaking}) \implies Innovation$ [MC, HAC]
  15. $speed(innovation) \propto \frac {1}{\text{half-life}(\text{mediocre projects})}$ [CCC, DD]
  16. $\Delta (Assumptions) 10\%> Innovation > \Delta (Assumptions) 1\%$ [NMN, SCP]
  17. Great Result/Lame Result/Suggest Go/1.0 Bonus/0.0 Bonus/Suggest No/Go 1.1 bonus/1.1 bonus [AKRD, MAR, TS]

Innovation is easy to be "for" but there's not much help out there for concrete ways to make it happen. Of course, we start with what we mean by "innovation." First, it's not the same as "improvement." Innovation is process is coming up with something that's enough better that you would not get there via incremental improvements.

Being innovative is not something you can just do when you need it. More a something you are. "Not being innovative is also habit forming."

$100 \times Innovation(1) \gg 1 \times Innovation(100)$

(15:30) We read this as "One hundred times a small innovation bet yields more than 1 instance of a big innovation bet."

To be fair, we might rewrite this as $E(100 \times Innovation(1)) \gg E(Innovation(100))$, that is, the expected value of 100 small-bet innovations is much bigger than the expected value of a single innovation bet 100 times as large.

Small bets vs. big bets. Case example: pottery class grade based on quality of final project vs. quantity of clay you use. How do you create an environment that promotes these small bets? Require that you only get my attention on an innovative idea if you bring me 10 other things that you have decided are not good ideas. You force folks to not get emotionally attached to one idea. You have to have actually prototyped or implemented things that you threw out.

$Innovation(Someone) \gg Innovation (everyone) \gg Innovation (no one)$

(4:30)"Innovation is not the kind of thing you can put in a drawer and just take out when you need it…it's a cultural thing, it's a habit. Innovation is habit forming and not being innovative is also habit forming."

The amount of innovation you get when someone is responsible for it is more than you get when "everyone" is innovative which is better than no one being charged with being innovative.

You get more innovation when you have a chief innovation officer. Not how smart she is. It has to be weird. It has to be counter-intuitive. It has to be upsetting even to the people in the organization.
Of course if you have no one thinking their job is to be innovative. But the most common scenario - everyone is innovative - is not much better. If everyone is responsible, no one is accountable. You need to have someone who is responsible for driving innovation. Not the same as CTO. It is an explicit role. Also, don't make the mistake of appointing the person who once did some particular innovation. It's about perspective shifting so what you want is to select people on the capacity to perspective shift and to get others to perspective shift.

$!Story \implies !Innovation$

"If you don't have a story, you don't have innovation." (8:10--11:05)

You can say "I invented X" but it's not useful until you talk about the benefits. For whom? Make the movie poster. Write the press release. If you tell your folks that you won't listen to any of their ideas unless they have a story, then what happens is a bunch of the stuff they consider bringing to you they just won't because there is no story or they aren't excited enough about it to make one up. But when they do come to you, they'll have spent time coming up with a story, be able to communicate the WHY of their enthusiasm, and by virtue of them thinking up the story that they will tell you they have done a good bit of the work that you would have to do to evaluate their idea. They come with that conversation already underway. And if you like it, you have the first draft of the story that you will take to the next person.

Innovation is about perspective shifting.

$Innovation(Story) = Benefit(vision) + Product(vision) + How(vision) + WhyNow(vision)$

(55:30-) "We have to have a secret weapon and you need to tell me what that secret weapon is as a part of this…." "If you want to tell me about how we are going to change the world, and I'm not really interested if we are not going to change the world, you have to tell me about a problem .. where the benefit would be that the problem goes away."

Continuing from material on needing to have a story. A feature is not an innovation. Get them to make the movie poster. Tell us what the ad sounds like. Show us the brochure.

Product that "releases the value"
How are we going to do this: the nugget.
WhyNow: Why didn't it already happen? Can't just be that we are smarter or have more money. We need to have a secret weapon.

$!Understand Details + !Delegate \implies !Innovation$


Take the time to understand what the innovation is and trust the innovator or create good metrics. Not doing so relies on your (supervisor's) intuition, but it's not intuitive. Spend the time to get over the counter-intuitive threshold. "Great, here's some resources, here's what I'd like to see you do…." Otherwise, no innovation or you are just betting on your intuition (which is by definition inside not outside

$Effort(Innovation_{V2.0}) < Effort(Innovation_{V1.2})$

([46:30-~50:00) "The effort to create version 2.0 of something, anything, that is innovation as opposed to improvement, is often less than the cost of making the improvement."

If you ask folks to improve something by 10% they will sweat it for weeks, zeroing in on each little detail and part and maybe eek out 8%. On the other hand, if you say you want it 300% better, they'll scrap the current model and try something completely new.

$P(Innovation \mid Bob) > P(Innovation \mid Idea_Z)$

(1:07:00-1:09:00) In the VC world they say invest in the person, not the idea, but often enough this is not what happens inside of organizations.

Consider how we kill this when the very folks who got the company going come up with a "company killing" idea and then

$\text{Idea Transplant} \implies Innovation$

(57:50-1:00:12, 58:20) "the easiest way to find a technology shift…" "an instance from a class, not a new class; innovation is about building a new class"

Take away here is to distinguish idea transplant from "me too." The latter is when you say "Facebook for dogs;" the former is when you wonder if there might be a way to use page rank logic to identify investment opportunities.

$\text{New Tools} + \text{New Metrics} \implies Innovation$

"(20:00) We haven't invested in the tools." No reward for building or developing new tools. Bacon quote: “If we are to achieve things never before accomplished we must employ methods never before attempted.” Every time you bring me an innovation, bring with it the main tool that helped you get to this idea. And I want an idea for how to make that tool better.

Around 1:03:00 he talks about what the boss and innovator need to do - set up milestones that innovator will need to knock down as part of the innovation process. This is in connection with question about how do you get a boss to be interested in a good idea.

$CleanUp(\text{painful workarounds} + \text{rule breaking}) \implies Innovation$

"One of the best places to look for innovation is where people are torturing themselves." (1:03:35-1:06:20)

Harvest weirdness.

If there are folks out there doing "bad" things with your product then you are being hit over the head with a message: that is your innovation.

Mentioned around 1:01:00
In connection with rule breaking: if everyone thinks you're wrong, you probably are. if everyone thinks you are right, then you are late. If 3% of smart folks think you are right and the rest think it's the stupidest idea then you are on to something.

$speed(innovation) \propto \frac {1}{\text{half-life}(\text{mediocre projects})}$

(52:30-54:15) "The speed of innovation is inversely proportional to the half-life of mediocre projects. … Most of [the cool ideas] are terrible … and unless you are Steve Jobs … you can't just look at one of these ideas and know whether it's going to be a billion dollar a year thing. … And so the rate-limiting step in innovation is not coming up with ideas, it is killing projects." "We won something when we kill a project."

We read this as "the speed of innovation is inversely proportional to the half-life of mediocre projects. The half-life means how long it takes for half of all the mediocre projects to disappear - it's a measure of how long we tend to be stuck with bad ideas. The inverse means that the longer this is, the less innovation you get - the bigger the denominator on the right, the smaller this term is.

The first reason why this is so is pretty obvious.

The hard part of innovation is NOT coming up with radically cool ideas. But most cool ideas are terrible. It's a little counter-intuitive - that we want to slide the idea killing bar over in the direction of good because bad ideas cost so much.

$Rebuild = Innovate$

"Rebuilding actually causes innovation." (27:22-29:12, 47:45)

You don't make it the same the second time. It's better. When we ask folks to rebuild something we let people start from where they were. How about rewarding folks not for how many lines of code they write but how many lines they delete. You can always add 1000 lines of gibberish to something, but try deleting 1000 lines from something that already works. You have to really understand the thing. Force people to start over. You don't just save yourself from the bad parts of lock-in. It actually drives innovation.

We don't ask for big enough when we ask for innovation.

$\text{Technology Judo} \gg \text{Technology Kung Fu}$

( 29:20-33:50, 1:09:00) "Do you want a vase or do you want a mechanism for displaying flowers?"

Picking the right problem.
The point here is that Kung Fu is a power game where as Judo is where you use the strength and moves of your opponent. With regard to engineering, the point is that the engineer does not pick the problem - they say, "you tell me what you want to do and I'll figure out how to do it." Judo is perspective shift. Teller example: voice over IP.

Solving the problem vs. changing the problem. Entrepreneur, designer, marketer: point is it is not part of the training of an engineer. This is true of academics and academic administrators in the extreme.

Technology ends up NOT being the hard part.

Ask explicitly, "how can we perspective shift so that the technology problem goes away?"

$\Delta(Assumptions) \implies Innovation$

(58:10) "…allows you to break the right assumptions…" (connection with idea transplant)

$\Delta (Assumptions) 10\%> Innovation > \Delta (Assumptions) 1\%$

(35:40) "Let's say there are 100 basic assumptions in some field … and innovation is picking some of those assumptions and breaking them. It also involves religiously following the best practices in the areas in which you are not actively breaking the assumptions."

We would read this equation as saying "innovation is in between the place where you throw out 10% of the assumptions and the place where you only throw out 1% of the assumptions." The former is too much, the latter too little. Assumptions are sacred to the experts - if you want to break them all, the experts won't hear you. So one thing it means is that you can't have just experts vetting ideas. Another thing is you should seek those few experts who have some iconoclast character because they'll entertain the breaking of assumptions.

Those kind of experts are especially useful because they'll go on the journey with you and push you hard every step of the way.

Note: in academe, as elsewhere, you cannot get this from consultants because there stock in trade is to give you what the experts assume. They are assumption and expertise peddlers par excellence.

$\text{Milestones, Bonuses, and Project Cancellation}$

(40:30-43:00) "The cost of dumb projects running forward is huge relative to those bonuses."


You have a group working on some crazy idea and you agree on a milestone and then you say "if you meet this milestone I'll give you a bonus X, if you miss the milestone you get nothing, but if you agree to kill the project now I will give you the bonus plus 10%."

It looks like something people would "game" - that is, dishonestly take advantage of, but they don't. In fact, what you have is an experiment that let's you detect the value of emotional attachment to projects - if there is some multiplier at which folks would abandon.

Important thing is that the cost of the bonus is so much less than the cost of running the experiment - continuing on the project in the hope that it pays off.

This connects to organizational psychology. If we want to be turning down most ideas then the folks with ideas have to have very thick skin. You have to figure out ways to really - symbolically and materially - reward them for cancelling projects.

And the giant value here is that you are tapping into the minds of the people who deep down really understand the promise of this idea.

Make the technology problem go away. Technology is not the hard part.


Other Observations

  1. Innovation and quality might be at odds with one another. The latter is about regularity and standardization of things already in the market. Often enough you have to relax constraints to get innovation (though sometimes you can generate innovation
  2. Innovation is by definition counter-intuitive
  3. (34:40)"You need to isolate the innovators from people who have an emotional or a practical stake in the innovators being wrong."


Add a New Comment