Bell Labs was only profitable in conjunction with AT+T

Bell Labs did pay for itself by making “The System” of AT&T more efficient. Xerox PARC did as well via the Laser Printer. However, most corporate research labs have more tenuous arguments for their profitability. Note that in both cases, the profitable outputs were not the most impactful outputs. There is a significant class of innovations that would create drastically less value for the world if their value had been captured by their creators. This disjoint between profitable outputs and outputs later hailed as valuable and important happens so often that I can’t believe it’s a coincidence. Seen vs Unseen.

It’s worth paying attention to the mechanism by which both Bell Labs and PARC (and extrapolating wildly, other corporate labs) pay for themselves. In both cases profit was through leveraging the scale and existing competence of the parent company. In Bell Labs’ case, it was by simply reducing cost at scale and in PARCs case it was something that could be produced at sold at scale with the same factories and sales channels.

The above quote is also notable because it illustrates how the efficiency gains were actually existential threats to AT+T because of the roundabout logic of “monopoly maintenance requires continuous improvement and expansion which would be impossible without finding a lead replacement.” Alignment requires existential threats

There’s a strong and a weak proposition here. The strong proposition is that without AT+T’s scale, Bell Labs would not be profitable. The weak proposition is that even given AT+T’s scale, the transaction costs would be too high for Bell Labs to have been profitable if it were a fully independent organization.

The strong proposition is pretty straightforward - any invention that isn’t wildly expensive like early computers and rocket ships requires scale to create value. To get scale, non-software inventions require either large manufacturing and sales apparatus and/or a large system to deploy on (in the case of efficiency-type inventions^1.) So, the profitable inventions from Bell Labs leaned on AT+T’s existing scale.

The weak proposition is more debatable but important to consider. In a world with zero Transaction costs, it wouldn’t matter whether Bell Labs was joined at the hip to AT+T - if there was an invention that would save AT+T billions of dollars, it would pay a large fraction of that gladly and Bell Labs would be able to support itself as a fully separate institution. However, transaction costs are real (see: The Nature of the Firm.) An independent Bell Labs need to convince a string of non-technical executives at AT+T that they would eventually run out of lead, the size of the problem, and the fact that the synthetic material would work as well as tried-and-true lead. On top of that, Bell Labs would need to realize that they should develop a lead replacement or what specific affordances need to be which is the sort of insight that is non-obvious if you aren’t constantly talking to the people who manufacture the cables. Prototyping needs manufacturing in the room. On top of that, you would then need to convince someone to fund development for the lead-sheath replacement (either internally to your org, for whom cable sheathing is not an existential concern, or externally) which gets you into a chicken-and-egg situation with all the previous considerations.

There’s a fuzzy thought here about value capture depending on where you draw a box around the system.


^1:Efficiency type heuretics are especially dependent on large systems because of Transaction costs. If an efficiency improvement makes something $1 more efficient per node, even if there are a billion nodes in the world, if each one is in a separate system and you need to sell to each one individually the transaction costs make it not worthwhile to create and sell it. However, if those billion nodes are concentrated in a few systems, its well worth it.

Web URL for this note

Comment on this note