Recently a few of us were discussing about what should be the focus of a new company in analytics space - products or services? The question was not purely theoretical as some members of the group are in process of launching a new analytics company. In this group, we had 3 kinds of people, technology experts, predictive analytics experts and someone who has built and sold analytics companies before. Naturally, I expected that we would have difference of opinion given the varied nature of experience and roles in the team. Surprisingly, though, we all felt strongly that for an analytics company to succeed pure services model will not work. Our reasons for same were as follows:
First is the IT organization of the enterprise - IT organization is actually very used to services model. They work on projects and it's a practice in IT world to hire contractors against project dollars. In addition, the coding portion of IT work renders itself a bit more easily to decoupling and hence the offshoring to cheaper locations like India becomes feasible.
Business teams on the other hand are very wary of services model. First, they usually get payroll dollars and not project dollars. Using payroll dollars they can only hire company staff and hiring contractors requires additional approvals. The project dollars that the business has are usually for one time needs (for instance, customer segmentation) and these get outsourced as a project to an established vendor in that niche space.
Even when the enterprise signs up for using offshore teams through captive centers or vendors, they are very concerned about lack of business knowledge of these teams. And they are usually right about it - it does take months to transfer the specifics of the business that are so important for providing appropriate analytic solutions.
For these very reasons, I have found business teams more open to buying a product to meet their specific need instead of investing time and effort in transferring business knowledge to another team. IT teams on the other hand are vary of buying analytic products unless they are sure that business would use these - else what's the point in adding another product to their list of things to maintain and support?
In summary, if your services or solutions are targeted at business users, its better to productize them to a certain extent and complement the same with service offerings for higher end predictive analytics that requires higher degree of customization.
- Services only model is not easily scalable as its very people dependent. Requires lot of hiring, training etc. as new deals get signed.
- Analytics in today's world is not used by a privileged few. In fact, its power can be truly unleashed only when everyone in the enterprise can access it and use it to make more informed, timely decisions. And products that can be used by business teams are the only way to democratize analytics
- The sheer size and kind of data available today makes it dependent on technology (hence products) to not just store it but to even use it.
- Last, but not the least, even the services side of analytics has evolved over the years. I remember, when I joined analytics as a profession, predictive modelers were held in very high esteem. Statistical modeling was as much about art as it was about science. Over the last 15 years I have seen that high end human modelers are being replaced by model building software by FICO or KXEN or SAS. A few months back I was introduced to SAS Forecast Server and its forecasting module. I was impressed at how hundreds of good quality Time Series forecasts can be generated by a team of only one or two statisticians. I am not saying that we don't need modelers but the fact is that we don't need as many and that has been made possible only by analytic products.
First is the IT organization of the enterprise - IT organization is actually very used to services model. They work on projects and it's a practice in IT world to hire contractors against project dollars. In addition, the coding portion of IT work renders itself a bit more easily to decoupling and hence the offshoring to cheaper locations like India becomes feasible.
Business teams on the other hand are very wary of services model. First, they usually get payroll dollars and not project dollars. Using payroll dollars they can only hire company staff and hiring contractors requires additional approvals. The project dollars that the business has are usually for one time needs (for instance, customer segmentation) and these get outsourced as a project to an established vendor in that niche space.
Even when the enterprise signs up for using offshore teams through captive centers or vendors, they are very concerned about lack of business knowledge of these teams. And they are usually right about it - it does take months to transfer the specifics of the business that are so important for providing appropriate analytic solutions.
For these very reasons, I have found business teams more open to buying a product to meet their specific need instead of investing time and effort in transferring business knowledge to another team. IT teams on the other hand are vary of buying analytic products unless they are sure that business would use these - else what's the point in adding another product to their list of things to maintain and support?
In summary, if your services or solutions are targeted at business users, its better to productize them to a certain extent and complement the same with service offerings for higher end predictive analytics that requires higher degree of customization.
No comments:
Post a Comment