Small and Medium Enterprises are now able to access tools that allow them to leverage data to make deep insights into the multiple facets of their operations. Machine learning-as-a-service is beginning to rise up, which can be attributed to vendors such as Amazon, Microsoft and Google.
These vendors provide an opportunity for companies to dip their toes into using analytics in their day to day operations and thus increase their efficiency. Moreover, they offer easy to use APIs and platforms that can run basic ML operations without the need for specialised hires. While this is good for a company that wishes to see the benefits of using ML to derive insights, a data-driven approach to the service itself is recommended.
Sign up for your weekly dose of what's up in emerging technology.
A well-scaled data science operation can provide insights into every part of the company’s architecture, and is vastly more efficient than siloed data or running an ML instance on a cloud service. However, an SME getting into using data is left with doubts as to how to go about integrating data into their processes. This article will take a look into how SMEs can enter the data science field with hiring after using ML as-a-service.
After a primary test with the ML as-a-service to see to what degree analytics can be used by the company, it is important to decide the approach to be taken to building a team. In startups, the analytics capabilities and requirements can be fulfilled by a variety of ready to use ML models and machines offered by cloud vendors. However, as data collection methods and analytics needs grow, companies are usually required to build models of their own and make hires to create, maintain and improve them.
It is important to note that existing sources of IT manpower can be leveraged to create a reshaped company structure. While a few new hires might be required, it is a much more economic and employee-friendly approach to re-skill and up-skill them. This would not only increase the value employees can provide, but can also make the company’s processes more efficient.
After acquiring analytics skills with easy to use services, a sufficient number of employees can be trained to derive data-driven insights to make their processes efficient individually.
Scaling For Deeper insights
Hires would have to be made for data scientists and engineers so that an approachable machine learning platform can be developed. This will not only enable a better adoption among the workforce, but also enable further scaling and deeper insights. Not to mention that such a personalised program will provide company-specific insights to the behavior of customers.
Companies have often made the transition from technology firms to analytics-driven juggernauts. In order to derive even deeper insights from your customers, data collection operations can also be upscaled in a variety of methods. This not only increases knowledge on the customer or the market, but also serves to better the algorithms used to derive insights.
Once the organisation has sufficiently scaled to an extent where analytics becomes the main driving force behind a majority of the operations, a different approach needs to be taken.
Organisation-Wide Scaling With CoE
A Center of Excellence approach is usually the one that bigger companies take, so that they can keep centralised power within the corporate center. This translates to a hub and spoke model, where manpower is allocated to different units in the organisation while sharing data and resources. This will also enable a higher degree of coordination between resources, and focus will remain on important business centers.
At this point, the company should spend less time hiring people for each title and focus on understanding what roles one individual data specialist can fulfill. This will lead to a generalist style of employees that can be deployed for any issue regarding analytics in the company. Mainly, data scientists who can work on large datasets and who understand the theory behind the science are required. They must also understand the architecture, infrastructure, and distributed programming of the analytics being used.
A company that has used this to great effect is AB InBev, one of the world’s largest distributors of alcohol. Ravinder K Sharma, Global Senior Director of Analytics Insights at the company, has been instrumental in establishing an Analytics CoE in India with 150 data scientists recruited over the last two years. This has allowed them to become an analytics-driven company, building analytics solutions at scale using a hub and spoke model.
The analytics CoE is compared to an ‘in-house analytics firm’ which functions in various domains. This includes pricing domain for revenue management, consumer sensitivity to price demand, promo performance, a multidimensional forecasting ability that predicts industry beer volumes and brewery level SKU volume to manage demand.
The Bottom Line
Analytics are one of the most important tools to ensure the scalable and organic growth of a company. It is important to gain insights from consumer-driven data to ensure that the company makes smart decisions to increase their presence in the market.