Make it cheap to be wrong

Release Rollbacks, Project Failures etc are biggest night mares for any IT project teams. Imagine the amount of pressure and stress it puts on everyone on the team to be perfect and to be correct every time anything is released. What does this create? What is the adverse effect of being perfect all the time? - If you haven't guessed it till now - It is Innovation - It is the ability to experiment which gets hidden behind these huge hyped pressures.

Organizations and Startup's who wish to Grow need to take a call and make it cheap to be wrong. It takes time and effort to make it possible but the results out of it are definitely worth it for sustaining growth. Having looked at multiple scenarios I personally reckon few methods that should be followed to make the whole process of IT development and innovation exciting and fueled with quick and validated learning (customer feedback).

Split-Test Experiment

Split-Test experiment also known as A/B testing is one in which different versions of a product are offered to the customers at the same time. If you look at this closely this is a great tool and gives every innovator and developer liberty and freedom to try things out with out going horribly wrong. Even the craziest of ideas can get validated through this without impacting all of your customers or the brand. Usually no leader would let a crazy idea go out to the market.

[Please note - It is definitely an accountability of the leaders and owners to define the percentage of split, and what functional flows that can be open for split test as part of the innovation sandbox]

There is also a big effort involved in measuring the success of split tests and understanding customer feedback but definitely a cheaper way to test and recover if something goes really wrong (i.e. customer does not like it)

Batch

Keep the batch size as small as possible also known as single piece flow in lean manufacturing. Batch size refers to how much work moves from one stage to the next at a time. Take an example of hundred letters which need to be folded, stuffed in an envelope and stamped. Intuitive way of doing it is folding hundred letters at a time, then closing them all in envelopes and them stamping them all together. Here the batch size is 100. But lean, faster and effective way of doing is stuffing and stamping one envelope at a time with a batch size of 1. [Don't agree?]

If you observe if you fold hundred together it seems more efficient to repeat the same take over and over and improve speed but in process oriented work like this individual performance is not nearly as important as the overall performance of the system. But more importantly what if the envelopes are of small size??? In the large batch approach you will not realize that until you have folded all the letters and started the stuffing process. But in a single piece flow you will realize it much earlier and no rework too. Quick customer feedback and remember quick finished products every second.

In short keep the batch size low and deliver more often. It is again cheap to recover if something goes wrong.

Adaptive Ecosystem

My third and an important step is to built a robust and friendly ecosystem to move ideas and code from the developers mind to the customers hand. Be it a prioritized product backlog or a development operations model or automated testing the need of the hour is to make it fast and fool proof. Open source community has given so many tools which can be used these days to build once and deploy everywhere which makes it cheap to try new stuff. The ecosystem need to be built which removes the dependency of huge release and deployment plans and gives the power in the hands of the developer to experiment thus removing the huge costs involved with releases.

Above are some of my thoughts in making it cheap to be wrong. Above all of these what is required is the passion and steer from organization leadership. The leadership need to understand that if they want the company to grow and be an adaptive organization then the traditional ways have to go away. They need to provide scarce but secure resources, Independent development authority and a personal stake in the outcome.

Lets make it CHEAP to be WRONG and I am sure results will be different.