The Use of Agile Methods by the Entrepreneur

On Tuesday, Israel Gat and Sebastian Hassinger gave a talk at Agile Austin on The Use of Agile Methods by the Entrepreneur.

Israel started by talking about market disruptions and the synergy that software brings to them (a follow up to a recent blog entry he wrote on the subject):

• Major disruptions have historically occurred infrequently (every 50 years or so): mills, railways, steel, automobiles, microprocessors. The last of which came along in 1971. What's next?
• Some theorize that disruptions are no longer the exception, but are now the rule.
• You can try to sense these and adapt to them to capitalize or you can try to change the market yourself.
• Software is a good fit due to its maleability.

He then went into some of the issues that software brings to the table:

• All known compound objects decay and become more complex with the passage of time unless effort is exerted to maintain them (Capers Jones). For software, this is 3-5 years.
• Technical debt in software complicates estimations. I.e., if you don't keep it in line, your estimates will become bigger and bigger over time as it becomes more difficult to make change.
• See their presentation for a good chart showing that technical debt is the difference between the actual cost of change and the ideal cost of change.

Sebastian then talked about agile contracts. In particular, how to structure contracts to facilitate and capitalize on agility.

• Trust is a major issue at the beginning of the relationship.
• Fixed price is the most common approach since it isolates the customer from the risk of overruns. Agile groups can increase margins with this approach since productivity is typically higher. It doesn't promote change though.
• Time and materials is compatible, but doesn't encourage the development shop to be disciplined.
• Incremental payment for incremental delivery (paying by iteration) requires trust by the customer since it doesn't spell out the end results that they need.
• Fixed price per story point (or function point) is a hard sell. It requires work up front and trust (since estimates are in abstract units and can be manipulated).
• Most intriguing is Sutherland's proposal: "Money for Nothing and Change For Free". It is essentially a fixed price contract with two additional clauses (executed at the customer's discretion). The first is that they can cancel early (since highest value is delivered first) for a 20% penalty (best if the percentage aligns with the vendor's profit margin as this prevents bias). The second is that the customer can make changes for no cost as long as they take out stories with an equivalent number of points.
• Customers don't understand that 80% may be enough.
• The more experienced a customer is, the less likely they are to be willing to experiment with new approaches.
• Once a customer sees the final number, it is "stuck".

Sebastian than talked about how agility complements entrepreneurship:

• Most new businesses don't understand their customer / market. Agile helps to evolve the product to what it needs to be.
• It used to be that startups were done by technical folks. They started with a neat idea and tried to find a market for it. It is starting to become such that business folks can start the business without significant technical skills. In this case, the business design and product evolve together.
• Agile is more attractive to VC's as a result. Though many are skipping the VC stage as it is easier and easier to start a business without significant funding.

Overall, it was an interesting talk. It looked at agile from a different point of view than typically done.