Why Scale Matters

Jeff Jarvis says that scale doesn’t scale. Fred Wil­son fol­lows up by talk­ing about the new scale. While I agree that new mar­kets can be born out of aggre­ga­tion, I dis­agree with the con­cept of scale not scaling.

The prob­lem is one of, well, scale. Let’s say that I want to make a movie that will rival Star Wars in terms of spe­cial effects and will have a plot that will take the story all over the world. So I can look at the aggre­gated model to make that movie. I’ll put the word out on my site and, hope­fully, peo­ple will con­tribute money and/or ser­vices. Granted, this process will take time as I’m not a movie insider and have no proven track record in terms of mak­ing movies. So, I now have to limit my ambi­tion and make a short demo.

I do that and it’s good enough to con­vince some peo­ple but not good enough to con­vince enough peo­ple to raise the funds. Let’s assume, for argument’s sake that I man­age to aggre­gate 10,000 peo­ple and they are all will­ing to chip in $100 towards mak­ing the giant movie. That’s really great, I now have a fund of $1 mil­lion. How­ever, the ambi­tious project needs 10 times that. What do I do then?

In the scale world, one of the peo­ple who sees the demo is a devel­op­ment per­son for a major stu­dio. They like the demo, believe that there’s an audi­ence for the story and throw $10 mil­lion behind the project. Some­how, I man­age to get this project done on that bud­get and have a com­pleted movie. The next thing is mar­ket­ing it… Gen­er­ally it takes a lot of money to mar­ket but, there’s one thing: if you’re with a movie stu­dio, it’s likely that the stu­dio already has pre-established rela­tion­ships with media and can get me inter­views, which help pro­mote the prod­uct. It’s also very pos­si­ble, in these days of media con­sol­i­da­tion, that the stu­dio which spon­sors my movie actu­ally owns some other media and can use those for cross-promotional effort, there­fore low­er­ing the cost of marketing.

Granted, I used a movie as an exam­ple but what about retail. I want to get a prod­uct out. This means design­ing the prod­uct, putting it together, mar­ket­ing it, and sell­ing it. I could aggre­gate those func­tions (and in many cases, it is being done) but it depends on the mar­ket. For exam­ple, if I wanted to intro­duce a new com­puter, would I be more suc­cess­ful as an inde­pen­dent player start­ing from the ground up or as a player within an estab­lished orga­ni­za­tion (let’s say Dell or Apple, for exam­ple) which already has all the parts?

Going beyond com­put­ing, would I be more suc­cess­ful intro­duc­ing a new cereal as an inde­pen­dent firm, aggre­gat­ing rela­tion­ships across sev­eral indus­tries, or as some­one work­ing for a large scale player, which already has all the parts?

Telecom­mu­ni­ca­tion actu­ally works within the model Jeff pro­poses but more because it is a sin­gle service/technology built on top of other tech (ie. the Inter­net infra­struc­ture is enough to sup­port it and has been built and VoIP can now ride as an extra on a multi-use network)

Labor may look like a per­fect model for aggre­ga­tion but the main­te­nance part and track­ing needs to be empha­sized here. Also to be taken into account is the level of expe­ri­ence of the labor­ers. For exam­ple, I can aggre­gate and off­shore an IT project but I can’t ask the same peo­ple to do my taxes or build a car. Why? Because of a skill gap. Off­shoring is inter­est­ing but, in every off­shoring project I’ve heard of or have been involved in, one of the key suc­cess or fail­ure fac­tor has been skills and train­ing… and on a small project, aggre­ga­tion is more expen­sive than ded­i­cated workers.

Pol­i­tics could be a place where the model works but that one will depend on which coun­try you are deal­ing with. A good exam­ple of the fail­ure of an aggre­gate model was the French pres­i­den­tial elec­tion, where the left decided to present many can­di­dates and many mes­sages. The right, on the other hand, pre­sented two large-scale choices: a center-right can­di­date, and an extreme right one. In an open pri­mary model (where the top two can­di­dates are the one to go on to the next round), the right defeated the left, leav­ing the main event as a choice between center-right and extreme right.

In the US, because of the way most pri­maries are set (on a per-party basis), aggre­ga­tion doesn’t work as well as scale. The last pres­i­den­tial cam­paign showed two large scale orga­ni­za­tions (Democ­rats vs. Repub­li­can) and a bil­lion small scale groups that could aggre­gate (or not). 99+ per­cent of the votes went to the large scale orga­ni­za­tions, which both offered a dif­fer­ent “one-size-fits-all” message.

I believe there is some truth to the aggre­ga­tion model for non-tangible assets indus­tries. For exam­ple, in the dig­i­tal assets world, aggre­ga­tion makes sense. If your prod­uct has already been built and is only need­ing dis­tri­b­u­tion, aggre­ga­tion works (one of my con­cerns around the long tail is that it doesn’t seem to take into account the cost of pro­duc­tion and effi­ciency of return on invest­ment), then scale may not mat­ter as much but, in a lot of case, scale can accel­er­ate work because it does not require to take the extra step of actu­ally gen­er­at­ing and man­ag­ing aggre­gated mod­els. In a lot of way, scale is about pre-aggregate mod­els you can just plug into.