During my time at an artificial intelligence startup in Bangladesh, I pitched data solutions to enterprise clients ranging from NGOs to consumer goods firms to banks.
This experience helped me learn a lot about the state of data infrastructure in developing economies like Bangladesh. I learned that there were 3 predominant types of companies with regards to data culture in Bangladesh1.
Data Resistants
The first type I observed in Bangladesh was also the most prevalent - the data resistants. Data resistants don’t have data pipelines (the method of ingesting raw data and transporting it to a data store ready for analytics) or sophisticated dashboards - some data resistants don’t even use spreadsheets. Data resistants exist in every industry, from small to large and from private to public sectors.
Interestingly, data resistants aren’t just old companies that have been left behind - many are thriving behemoths. They range from large goods distributor networks to government organizations, charities, and some of the largest consumer goods firms in the country.
It's amusing to think that these massive firms (some of the largest in the entire country) contribute nothing to the numbers below:
One could argue that many data resistants do have non-digital data storage (receipts and paper records) and use that data to make business decisions. However, my argument here is: how reliable are data-driven decisions when your store of data is mid-level managers (in the sense that those paper records are analyzed by executives who transmit the findings to their managers)?
Data analytics is critical to business growth, and there’s proof to back it up - globally, businesses that use big data saw an average profit increase of 8%, and a 10% reduction in overall cost (BARC research). Approximately 40% of companies worldwide also use big data. Then why do these non-data-driven businesses struggle with using data, and how do they grow regardless?
Why don’t data resistants use data?
They’re operating in industries where it is the norm to not record or leverage data
This could be due to challenges in setting up data infrastructure. For example, some consumer goods companies struggle with setting up data recording solutions due to technologically handicapped store owners or field employees, making it difficult (sometimes impossible) to capture data at the source.
Another reason is that most companies do not want to challenge the status quo; these companies have thrived without data pipelines, so why are they needed now? There also exists a hierarchical leadership structure in most local firms where upper-level positions are filled by employees based on their age instead of merit - a common practice in Bangladesh. Family businesses are also common in Bangladesh, even among the largest conglomerates, where top management is occupied by older members of the family (who are more likely to stick to traditional business practices).
It’s human nature to resist change, especially for the older generation who are mostly out of touch with technology.
They have no “proper” competition and thus no incentive to use data to innovate
A basic business lesson is that a monopoly (a market where there is one large seller of a product) is bad for everyone except the monopoly itself (I’m simplifying, but it holds in most cases). Monopolies and oligopolies (markets shared by a small number of sellers) usually don’t have much reason to innovate.
While a lot of industries in Bangladesh aren't exactly monopolies or oligopolies, most are dominated by traditional local firms, which don’t provide “proper” competition. When I say “proper” competition - I mean competition that uses data to innovate their products and services (I'm not the best at naming things). So while competition exists in these industries, competitors do things the old way. Therefore, a lot of these companies have no incentive to use data and improve.
One reason for this lack of “proper” competition could be the first-mover advantage - many local firms have a foothold simply because they were the first, not necessarily because they were the best. Other reasons include high initial capital requirements, legal red tape, corruption and various other challenges in the entrepreneurial landscape, which make it harder for newer innovative firms to enter.
Bangladesh ranks 168th in the Ease of Doing Business Index out of 190 countries which highlights how hard it is to start a business in the country.
All this added with the fact that it’s expensive to build up data infrastructure (cloud or on-premises data storage, third-party/custom analytics solutions and a data team) means that there’s not a lot of incentive to innovate.
A hot new startup using data-driven decision-making to gain an advantage over (less data-savvy) competitors doesn't exist in the current Bangladeshi market.
How do data resistants grow?
You might imagine that data resistants would struggle to grow when they have such an incomplete view of their data. But the reality is that they still succeed due to a few factors. Most of these factors are external and unique to the market they operate in, so there's no secret sauce here to replicate. Sorry, non-data-driven CEOs, you'll have to use your data to grow your business.
One advantage of operating in a country like Bangladesh is the rapid population growth and large existing population.
This, coupled with a rising middle class and higher levels of disposable income, makes most of these companies very profitable and allows them to grow rapidly.
A leading group of companies (mostly in the consumer goods space) grew 4.58x in the last decade despite being a data resistant. This company and many others in this space typically change their product lines based on gut instinct and a general idea of the market. According to an ex-employee (a mid-level manager), upper-level managers would ask for national sales reports for certain products, brainstorm reasons why some products didn't succeed while others did (sometimes using consumer focus groups but most of the time just based on personal judgment), and then write up action points to improve/discard/promote those products.
The same employee notes that "one product - a biscuit - was doing really well in rural markets but not so much in the cities. The problem is that in order for a product to continue, it needs to hit a national sales target - there were no regional targets. Just because national product revenue was low, and that's how far data decisions went in the company."
There were no disadvantages to not looking at the segment data (in this case, very simple regional segmentation), but the company's decision-makers didn't do so because they had no urgency to innovate - the market was growing and no one else was using data to innovate.
It's easy to grow as a non-data-driven company when your competitors (if they exist) are also not using data. It's easy to grow when no one else can come in and disrupt the market as well.
What about the other types of data cultures?
There are two other types of data cultures that I've seen - Data Eager and Badly Data Driven.
Badly Data Driven companies do record data, and they use it - just not optimally. While it can be argued that no company in the world is possibly using data 100% optimally, these companies actually have data cultures where being data-driven leads to more harm than good!
Data Eagers are companies that have a lot of data and want to use it; upper management understands the importance of data and insights generated from it, but the company doesn't yet have a proper data team or a robust data infrastructure.
You’ll have to wait for future articles to find out more about these companies 😉. Until then, subscribe and get the Substack app to stay updated!
Want me to clarify something in this article or want to know how I came up with such bad names? Leave a comment.
Surely there’s exceptions?
I've been a bit unfair in my categorization above, as some companies (albeit a small number) actually do leverage data very well and are thriving because of it.
Notable standouts are telcos (telecommunications services), where the three top players generated revenues of USD 3.2 billion in 2020. While the largest telecom operator (Grameenphone) enjoys around 48% (46.39% as of Oct 2021) market share - there is still a culture of innovation and optimization i.e. they have “proper” competition.
Most of these companies have robust data pipelines, a data lake/warehouse, and an analytics tool. They use the incredible amounts of insights they generate from their data to gain an edge over the competition, squeeze out every last bit of user revenue, and reduce churn.
The new wave of startups also leverages data to great effect - with some of these companies employing a large number of data professionals. These operate in multiple areas from ride-sharing to edtech to e-commerce platforms.
Does this apply to other countries?
My categorization of data cultures into the 3 aforementioned types is only prevalent in certain emerging economies. For it to hold, the country must fulfill some or most of the following factors:
It's hard to start new businesses (legal barriers, corruption, lack of investment, etc.)
Most industries are dominated by local traditional businesses that rarely innovate
Infrastructure needed to capture and store data (internet services, telecommunications, hardware imports, etc.) is underdeveloped
Lack of data talent, possibly due to a lack of data education or lack of career opportunities
Primary (agriculture) and secondary (manufacturing) sectors still represent a large portion of the economy
Fun fact: this article mentioned data 75 times
Thank you for reading Slightly Interesting! See you next week.
Note that we’re concerned with large enterprises that have the resources required to be data driven.