A coalition is easy to announce and hard to join. India 2022 had a real ambition, a serious set of members and a problem I have seen many times since: a company that wanted in could not tell what it would be agreeing to on Monday morning. As Xynteo's chief digital officer I worked on the part of that problem that is solvable, which is the architecture.
The uncomfortable thing
Coalitions of this kind fail in a specific way. They convene, they publish an ambition, they hold an annual event, and two years later nobody can name a thing that exists because of them. Commitment is rarely the problem. The ambition is stated at a level of abstraction where no budget holder can act on it, so every partner waits for someone else to go first.
So the work was to make the ambition divisible into pieces small enough that one company could own one, and specific enough that owning one meant spending money and reporting a result. Making it more inspiring would have made it worse.
The structure
The coalition was cut into four impact tracks, each mapped to a development goal the government of India had already committed to publicly, so that the target was somebody else's number and could not be quietly moved:
- Skills, against the goal of vocational training for 500 million people by 2022, and carrying Project Disha, which aimed to help one million underprivileged women learn marketable skills and connect them to income.
- Energise, aiming to produce enough energy for the daily requirements of 10 million people, against the goal of universal access to round-the-clock power.
- Connectedness, aiming to improve the livelihoods of one million people by expanding internet access, against the goal of broadband to 600,000 villages.
- Sanitation and waste management, against the goal of an indoor toilet in every home.
Each track had a founding partner who owned it rather than sponsored it. Each track ran the same cycle: identify a commercial opportunity sitting inside a human problem, incubate a pilot, and only then talk about scale. And the coalition recruited what it called catalysts, employees seconded from partner organizations to run the pilots, which is the mechanism that keeps a coalition from becoming a secretariat.
The room, briefly
The executive committee was the chief executives of Hindustan Unilever, the IKEA Foundation, Shell, Technip and GE South Asia, with an advisory board that included Sam Pitroda and Naina Lal Kidwai, and support from the government of India up to the President. I mention it because it sets the altitude of the conversation and the tolerance for vagueness in it, which is zero. A room like that does not need to be inspired. It needs to be told precisely what it is being asked to sign, and it will decline politely and permanently if you cannot say.
What changed
The tracks converted. Within the skills track the coalition was developing pilots with IKEA Retail, Jindal Stainless, Jones Lang LaSalle, the Taj Group of Hotels, TATA Strive, Sodexo, DNV GL and Future Group. The energise track had projects around waste to energy, rural solar with Simpa Networks, and fly ash reuse. Those are named companies attached to named projects, which is a different object from a coalition with a manifesto.
The other outcome was narrative. There was finally one document that explained the whole thing in commercial language, without asking anyone to donate their way to an outcome, and a partner could read it and know which track they belonged in before the first meeting. That sounds like a small thing. It is the difference between a coalition that grows by introduction and one that grows by referral.
What I would do differently
I would have made the coalition measure its own cycle time. Everything was measured in partners recruited and pilots launched, which are counts of activity. The number that would have told us whether the machine worked is the interval between a partner joining and that partner having something running in the field. Nobody was accountable for that number, so nobody shortened it. I now put cycle time into the first page of anything I build, because a program that cannot see its own latency will absorb any amount of goodwill and return it slowly.