Chapter 28 of 29 · Organising a place for discovery
1 The bottom line
Sort each failure by its cause, give failed work a second look, and check whether outsiders or your own breakthroughs threaten you. Bell Labs missed the integrated circuit, lost a waveguide bet and watched its own work undercut its monopoly.
2 Key takeaways
- Name the type of failure before you decide what to fix. Missed opportunities, mistakes of perception and mistakes of judgment have different causes and different fixes.
Jon Gertner, Bell Labs: the lab missed the integrated circuit, bet on waveguides that glass fiber made obsolete, and launched a Picturephone nobody wanted.
- Review failed work whenever a new problem appears. A weakness that sank a product in one setting can stop mattering in another.
Jon Gertner, Bell Labs: the silicon solar cell was too costly for rural phones, but in space nothing else could recharge batteries.
- Do not let a leader's belief or your sunk cost close a direction. You can own every ingredient and still lose the invention.
Jon Gertner, Bell Labs: Jack Morton believed multi-component chips could never be reliable and pursued functional devices instead.
- Never assume your network is a moat rivals cannot copy. Widely licensed technology lets others build what you built.
Jon Gertner, Bell Labs: MCI built a cheaper microwave long-distance network without durable gear or local subsidies.
- Ask whether your own breakthroughs remove the conditions that protect you. Your best work can spread, get improved and return as competition.
Jon Gertner, Bell Labs: an analog network needed one entity to control many amplifiers.
Shannon's theorem was the mathematical basis for breaking up the Bell System.
Jon Gertner, Bell Labs
2.1 Sort each failure by its cause
| Missed opportunity | Mistake of perception | Mistake of judgment | |
|---|---|---|---|
| Diagnostic question | Had the parts, someone else assembled them? | Better answer came from elsewhere? | Product failed the need? |
| Bell Labs case | Integrated circuit (Noyce, Kilby) | Waveguides beaten by glass fiber | Picturephone nobody wanted |
| What to examine | Parts you held and who combined them | Who solved it and their core skills | Function, cost and timing separately |
A failure only teaches you something once you know which of three kinds it was.
- Ask first whether you had the ingredients and someone else assembled the breakthrough. That is a missed opportunity
Jon Gertner, Bell Labs.
- List the parts you already held and who put them together first.
- A Bell Labs insider said: "We had all the elements to make an integrated circuit... But nobody had the foresight except Noyce and Kilby."
- Ask whether a better solution from elsewhere eclipsed your bet. That is a mistake of perception
Jon Gertner, Bell Labs.
- Look at who produced the better answer and what skills they started from.
- Such breakthroughs often come from groups with different core skills and different needs.
- Ask whether the product failed the need. That is a mistake of judgment
Jon Gertner, Bell Labs.
- Check function, cost and timing separately, because each points to a different fix.
- A product that failed on cost for one use may still be the only option for another.
2.2 Give failed work a second life
Shelved work can become essential once the right problem shows up.
- Keep an inventory of shelved and failed components, and review it when a new idea appears.
Jon Gertner, Bell Labs
- Record each item with the reason it failed.
- Check the list before you start building something new.
- Joe Chaffee's 1937 feedback receiver was considered largely useless. It was revived for Echo and Telstar.
- Look for a problem where the old weakness stops mattering.
Jon Gertner, Bell Labs
- Name the weakness that sank the product.
- Search for a use where that weakness does not count against it.
- Old work can become essential once the right problem shows up.
2.3 Test the beliefs that close off a direction
A confident "never" and money already spent can both blind you to the better path.
- Treat a leader's "this can never work" as a hypothesis to test.
Jon Gertner, Bell Labs
- When a hard manufacturing problem blocks a better design, test whether it can be solved later.
- Make dissent safe.
- Kilby and Noyce bet that reliability could be fixed later. They built the integrated circuit on Bell Labs processes like diffusion and photolithography.
- Check whether money already spent is shaping how you judge alternatives.
Jon Gertner, Bell Labs
- Ask people with no investment in your approach to judge the alternatives.
- Bell Labs spent years and millions on hollow waveguides. Baker still championed them in a 1968 speech.
2.4 Watch for outsiders who can copy you
The threat often comes from groups with different skills, and your network will not stop them.
- Watch outsiders with different expertise and no sunk investment.
Jon Gertner, Bell Labs
- Track fields whose core skills differ from yours but whose needs overlap.
- Charles Kao argued for pure glass fiber in 1966. Corning announced very low loss fiber in September 1970.
- Do not treat your existing network as a moat.
Jon Gertner, Bell Labs
- If your technology is widely licensed, others can build what you built.
- The attitude to avoid: "You don't have to worry about this, because we have the network. No one else has the network."
- Plan for low-cost entrants before regulators force the issue.
Jon Gertner, Bell Labs
- Expect them to target your most profitable segment without your obligations.
- In 1971 the FCC ordered AT&T to connect MCI to its local switching centers.
2.5 Check whether your own breakthroughs undercut you
Your best work can remove the conditions that protect your position.
- Map how each innovation changes the economics of your market.
Jon Gertner, Bell Labs
- Assume your work can spread, get copied and improved, and return as competition.
- Shannon's digital information let any company code and decode messages.
- List what protects your position, then ask whether your latest work weakens it.
Jon Gertner, Bell Labs
- Plan for the result before the market delivers it.
- Cheap, freely licensed transistors meant any company could do that coding and decoding. That undercut the case for the monopoly.
Sources
Part of Organising a place for discovery in the growth library.