Dogfooding software products

Here is the old story of Principia Mathematica written by Newton, when he was responding to a question by Halley about planetary orbits,  he only responded that he had calculated it mathematically and are always elliptical and had also lost the proof. But he promised to Halley he will reproduce the proof and goes into one of his famous 2 years retirement to write the entire Principia . It does not end there, like right now, Royal Society as publishers of books were in tough times, refused to publish Principia on financial grounds, and even when Halley was personally sponsoring the printing, the society itself was actually paying part for Halley’s employment with one of its previous unsuccessful commercial books called “History of Fishes“.

It was custom (as Bill Bryson puts it) that Newton never really got paid anything nor did he sponsor any money to get his book published, if someone else was interested, onus is theirs, in this case unfortunately Halley. And the society gave a raw deal by giving away copies of unsold books.

Payment in bad products

History of Fishes

Now the diametrically opposite context.

In a team built software product

1. there is no one single owner for idea / product

2. there is no individual revenue possibility or expense for the team

3. get paid anyways for whatever be the product or its quality or the purpose

That in my opinion is a bad recipe for quality. So here is the idea, “dogfood” the product i.e. every product team will need to find a market on its own for the developed product internal/external. And part of the revenue is directly shared. There are of course different model possibilities and I will go with the easiest and justifiable.

Elements of this model as I see it in my own warped way

1. Paying in product licenses

If the product that is being built is so good, can you please agree to take part of your regular salary as product license, and depending on everyone’s contribution, we will split the numbers only up to a certain limit, and then it will simply become a long tail % fixed, say after expiry of warranty on first release.

2. Order of license sale and revenues.

With a deep hierarchy in development, I believe the developers and architects’ license go first for sale, then the scrum masters or delivery managers, and then the higher levels of management later. Revenue sharing is truly from what was built for either direct effort or supervisory.

3. Expenses for such sale

This is a tricky area as most marketing functions are centralized and of course larger branding expenses at company level. So un-bundled marketing services should be costed accordingly, such as running an ad words campaign on Google, buying sponsored posts on linked in, or event/networking costs.

If it does not sell, then only the license part of your salary will get affected, and life goes on as developers find meaningful opportunities with real revenue possibilities to work on.

In a way a market gets created internally for projects/products to work on, and there is definitely a chance to design your career from inside by being open to trying and learning. I guess the same thing can also be done for a service, but a little more complicated.


H1 2009 Round Up

Here is a 2009 first half year round up. Not much as I see it, but I could still be happy that I am a great filter.


A great pitch will look like this when built with popular social objects, in this case Hugh‘s cartoon base.
Also of note is the Grabowski ratio in action M/E costs.


  • Within the CoP context I think desigining such spaces will energize and commit more people to the cause of knowledge. It is better and possible to design spaces for social interactions in your current context than blabber about knowledge eco-systems in ppts. Thanks to Kim for the hat tip.
  • In addition to Stephen P. Borgatti and Jos´e-Luis Molina paper , Toward ethical guidelines for network research in organizations
    this post will help kick start action in SNA if you are thinking of conducting one. Thanks to Graham Durant Law.
  • Interesting interpretations on network types and underlying SNA Methods. Great example on how easily we can misread a network diagram from Connectedness.
  • Though many people adopt NPS for analysing customer satisfacion scores longitudinally, I am seeing NPS in the context of communities for the first time
  • As unusual as it can get, community based km at SAP

Social Media Examples


  • We have been hearing stories get you jobs, make people change, spend more, trust more etc, here is a nice example from an eBay sale.
  • The classic story archetype of J Campbell more movies this summer and the next and the next, you never get bored it is that good an archetype.

Is there any another way to allow people make sense than by narratives so variations can fit to context?

Knowledge Representation

I have this fetish for information visualization and flowing data has been my primary solace, here is Fusion Tables from Google. This I think is a direct take on ManyEyes from IBM.

Random KM

Yeah I hear this big picture missing thing a lot.
  • From Dave Snowden, lists facets on opportunity scenarios in recession times. I need to apply trends and its system evolution laws for articulating the opportunities within an industry. I have been using STC (Stands for Size, Time and Cost) from the TRIZ toolkit for quite a while now and it is very powerful for scenario thinking.
  • In a way organizational knowledge base system is very similar to these online giving markets. Giving part means the existence of trust both by the giver and taker.
    Great pointers on the success elements for such a market.