Monday, June 19, 2017

The importance of a minimum viable product in development interventions

Tim Harford's last book Messy, has this to say about Amazon's turbulent initial years,
Bezos combined a grandiose vision with the sketchiest understanding of how the vision could be achieved... Bezos was making big claims to his customers... but he didn't know how those promises were going to be kept. He trusted that they would figure something out. On eight have thoughts that these early weeks were a good time to pause and regroup, to concentrated on making sure Amazon was able to deliver on its early promises. But Bezos believed in seizing opportunities rather than pausing for breath. In Amazon's second week of business, he received an email from David Filo and Jerry Yang, the founders of Yahoo. Filo and Yang wanted to list Amazon on the Yahoo home page - would that be okay? Bezos's computer guru warned him that it would be like trying to sip from a fire hose. Bezos ignored him and accepted the offer from Yahoo... The workload was inhuman... Accepting the Yahoo offer within a fortnight of launch was characteristic of Bezos... 
In 1999, Bezos decided to start stocking kitchen equipment. In Amazon warehouses that had been designed to store, sort and dispatch books, naked carving knives were suddenly scything down the chutes and into the sorting machines. Meanwhile the company's database would be asking whether the knife was a hardback or a paperback. 1999 was also the year in which Amazon started stocking toys... When Christmas came... Amazon employees across the US bought Costco and Toys'R'Us inventory in bulk and drove it to the Amazon warehouses... The scramble was too much for Amazon systems... Products lost somewhere in the vast distribution centres would send Amazon's databases haywire. Unshipped orders clogged the chutes in the sorting facility, each blockage spawning half a dozen further delays. Internally, the company was on its knees: it launched a 'Save Santa' campaign and Amazon staff members were booked into hotels near warehouses (two to a room) and didn't go home for a fortnight. As Christmas passed, 40% of the toy inventory was unsold and probably unsellable... In the summer of 2000 came the dot-com bust... There was a real possibility that Amazon wouldn't see Christmas 2001... As Brad Stone writes, "Amazon survived through a combination of conviction, improvisation, and luck".  
Harford points to the concept of OODA, coined by US Airforce colonel named John Boyd, 
OODA stands for 'Observe-Orient-Decide-Act' - or, in plain English, working out what's happening, then responding... If you could make quick decisions, that was good. If you had a strong sense of what was going on around you, that was good too. 
I am inclined to argue that this applies with even greater relevance for public policy implementation. The conventional wisdom on the implementation of new program and projects is that the leadership plans every implementation step to the last detail, allocating responsibilities and putting in place monitoring mechanisms, and then goes ahead and the executes them to perfection. Unfortunately,  this approach is unlikely to work with any complex development interventions.

All such implementations require an OODA-type approach. Figure out a reasonably good initial implementation plan, a minimum viable product (MVP). Put in place a very good war room with competent people. Respond very quickly to emergent problems and refine the implementation plan. Iterate extensively with short feedback loops and improve the implementation plan as much as possible over as short a period of time as possible. 

I have written about why this is perhaps the most effective approach in the roll-out of large infrastructure projects like a new airport or a new urban transportation project. It applies just as well to new projects in most development sectors.

The GST roll-out is a very good example of a situation where the MVP approach can be adopted. Given the complexity of responses of various stakeholders and resultant uncertainty, the best that Governments at Centre and States can do is to roll-out the platform and keep the powder dry so as to respond to emergent problems. The success or otherwise of the roll-out will critically depend on the swiftness and competence of the response. 

No comments: