Date: 18-03-2023 at 15:30 PM
When I was reading "No Filter" by Sarah Frier, I was surprised by the fact that Instagram faced slowdowns in its user growth. And each time their growth team identified and discovered new user cohorts, they optimized the content and suggestions accordingly. Here is a brief story of Instagram's growth story: the initial 100 million users were purely based on targeting the young population in the age group of 15 to 25. However, as they started seeing a decline in that specific user group, the growth team shifted their focus to the mature user group: married housewives, in the age group 28 to 40, working professionals in the age group 26 to 35, and so on. Based on each demography, they adjusted their content suggestions, recommendations, explore, etc. For example, if the cohort base is housewives, in the age group 28 to 40, the content suggestions mostly revolve around cooking, parenting, home decoration, etc.
I would not lie; before reading this book, I used to think that most of these companies acquire users by running ads or creating and optimizing viral loops blindly. After all, doesn't everyone want to use Instagram? I was wrong. And hence we have been careful not to consider India's population in India-1, 2, and 3. Here's why.
One of the things we at Jile have been working on - identifying and spotting our 1.2 billion potential customers in the x-y coordinates under the periphery of our nation. This was one of our top priorities because, when we were trying to understand why companies have not been able to build products for these customers, one of the challenges is unsustainable distribution costs. Technically, distribution costs would never be unsustainable if customers could pay for their jobs to be done. However, India with its relatively lower GDP per capita, makes customer groups price-sensitive. Since the distribution cost and price of the products are directly proportional, making products available at an affordable cost means we must figure out the x-y coordinates of our customers at the lowest possible cost.
Like all of you, we started thinking in terms of India-1, 2, and 3. We measure India-1, 2, and 3 based on the GDP per capita: India-1, INR 50,000/month (5 Cr), India-2, INR 20,000/month (10 Cr), and India-3, INR 7,000/month (120 Cr). The problem with this distribution method is that it gives zero understanding of x-y coordinates. What I mean by this is that out of India's ~3.287 million spare km geography, where do these India-1, 2, and 3 belong? Maybe India-1 and India-2 belong to the urban areas and India-3 to rural areas. We were super confused when thinking about or planning our distribution cost and desperately looking for something that could have given us a better understanding. And in that desperation, we came across Niti Aayog's report.
I wouldn't say this is the perfect distribution of India; however, it is a bit nuanced and can be utilized to plan the distribution better.
This method is based on a database: "National Health Profile - 2019," in which they used this quintiles-based system to distribute India's healthcare infrastructure. And I absolutely loved the entire concept. The quintile-based system is based on the economic status of the Indian population-based in urban and rural areas. Quintile (Q5) indicates the highest and Quintile (Q1) the lowest.
What can we understand by looking at this table?
Urban India is more uneven than rural India in terms of income.
The Income of Urban (Q1, Q2, Q3, and Q4) is almost similar to Rural (Q2, Q3, Q4, and Q5)
If we leave aside the Urban (Q4 and Q5), the right product can motivate the rest of the quintile population to pay for their needs. This also means products and services targeted for Urban (Q4 and Q5) might not be suitable for the rest of the quintiles. :)
This indicates that the distinction is not between urban and rural populations but rather within the overall population of India. By understanding this, we at Jile have been able to plan and identify our customer groups. We have created a separate document that summarizes attributes related to each quintile.
For example, we have noticed that trust formation in India is concentrated and hyperlocal. Concentrated trust is useful for low-involvement products, but for high-involvement products, hyperlocal trust sources such as friends, family, or long-lasting stakeholders function far better. For instance, my father trusts the postman (who has been in his job for more than three decades) for his investment decisions more than me, poor me! This is a common pattern in India. Similarly, ASHA workers are often the first point of contact for healthcare issues, suggestions, and decisions.
Therefore, we have identified the stakeholders who control hyperlocal trust for a set of 200 to 1000 families listed under each quintile (urban and rural). This has helped us to precisely target our audience, eliminate the cognitive load, and generate trust. We are still a work in progress.
We have found: it is relatively easy and cost-effective to acquire users in India (for example, the per-install cost for us is INR 5.69, and the per-signup is INR 7.32). However, taking money out of their pocket for solving their problems demand an Indian style. The Good news: the deep penetration of the internet has allowed us to reach almost the entire population as a demand side within less than two years.
Some of the data points in the above table are outdated, and we are eagerly awaiting the release of new data. We believe the new data will give us a better understanding and its potential to reduce distribution costs, further.
This is one of the reasons why we excluded the top 10% of India's population. Even though we have to target our potential customers in quintiles, we don't have to worry about the price point because it would be affordable for populations that cater to any quintile, whether urban or rural.
As I write this essay, I am currently in Gurugram. I purposely took a bus from Patna because there is no better way to gain a deeper understanding of progress than through visual data through your eyes. I think it has been around 14 to 15 months since I last visited Gurugram, and the place looks almost the same. However, the new cities that I passed through to get to Gurugram, such as Hajipur, Muzaffarpur in Bihar and Lucknow and Agra in Uttar Pradesh, are completely different from what they were just 12 months ago. It's mind-boggling to me the pace of progress in these rural and urban lower-quintiles. One particular observation is that 95% of the billboards, branding materials, business names, etc., in Lucknow and Agra are in Hindi. I am sure this would be the case for other quintiles as well - the language would be different: Tamil, Telegu, Kannada etc.
While I am in Gurugram, I visited the Golf Court, and my mind quickly decided that it wasn't for me. I find it fascinating how my mind can differentiate what it wants to consider or not.
Disclaimer: The household data is based on 2020, but the income data is based on the 2011 census. Please do not get confused. We expect to have the new data within six to 12 months.
Also, I will be in Gurugram until March 22nd at 2:00 PM. If you think we can learn from each other, I would love to meet you!
Thanks for reading - if you find this essay helpful share this with your network! I will see you all the next week :)