Ask any small and medium-sized business (SMB) about its top priority, and typically the answer will revolve around retaining the growing customers and increasing revenue. Similarly, a majority would say that cost is the single most important obstacle in adopting technology.
It’s here that analytics, a hot topic today, is generally placed well to address the objective highlighted by such SMBs. But the overall adoptions of analytics is less than 19% globally. How can SMBs, with their cost constraints, be able to take advantage of analytics?
As with other technology implementations, organizations that wish to deploy predictive analytics i.e. data mining (answering the question: “What will happen”, and not just “what happened”), need to answer four fundamental questions: Which business areas will we deploy analytics for; what will be the benefit/returns; how will I execute; and how much will I have to invest.
Most organizations refrain themselves from implementing analytics solutions by assuming they do not have enough data. While profile and demographic data is hard to come by, transactional data is always available and reliable.
Macro data about markets/economy is also adequately available. Using these, analytics has been majorly deployed for application areas centred on increasing revenue.
For example, a data service provider with revenue of around Rs.50 crore is deploying analytics to predict churn for its key territories. A Rs.300 crore financial distribution company is deploying analytics to improve cross-sell to its customers through its over 180 branches. A mutual fund house with less than Rs.10,000 crore AUM (assets under management) is thinking of investing in analytics to improve its distributor productivity. A young insurance provider with less than 5 lakh policies is attempting to use analytics to understand the behaviour of its bank partners.
Which area should you focus on will depend on your priorities, the industry you belong to, and the data availability. “How will I execute” has two parts to it—people and processes. As long as the operational data (sales, bookings, etc.) is available in electronic form (enterprise resources planning, or ERP is nice, excel sheets are fine too), one is good to go. Again on the execution front, one could use excel for the output/feedback and email for output dissemination.
On the people front, predictive analytics does require certain level of data science and statistical skills. For an SMB its best to outsource these skills, as it may not always be possible to remunerate and provide a career graph that aligns with their individual aspirations.
The only thing that separates SMBs from large corporations is the quantum of investments they can make. Analytics has been generally perceived as “good for large corporations” due to its price perception. Traditionally the capex (capital expenditure) investment for analytical programmes ranges between Rs.1 crore and Rs.3 crore, depending on the scope, data, users, performance, etc.
This is similar to the challenge telecom operators faced in early 2000s with high handset costs and high tariffs leading to low adoption. When struggling to extend their consumption in non-urban areas, FMCG (fast moving consumer goods) companies innovated with sachets that brought about a sea change in consumption, like the small ticket prepaid transformed telecom adoption.
Going from capex (capital expenditure) to opex (operational expenditure) investment structure fuelled the growth in both these cases. Can something similar be done to enable SMBs adopt analytics?
The advent of Big Data over the past couple of years has brought along with it a new wave of technology platforms (developed by organizations such as Yahoo and Google). When released to the general public, these became available for everyone to use without any licence cost, and thus became a part of the open source family—read Hadoop, R and Linux.
Hadoop is now the default data platform for most large-scale businesses, including e-commerce firms. R is the data mining tool, similar to SAS or SPSS but without any licence costs. When combined with opex-friendly technologies such as the cloud, these rich open-source platforms can bring down the implementation cost of analytics projects by 40-80%.
With the four foundation questions discussed, are we good to go? Start with one business area, try before you buy, and iterate what works. SMBs, here’s an opportunity that will help you get revenue, retain customers—all at an affordable cost.
This article was first published in the Mint on 27th Nov 2015
