How about a ‘managed decision science’ program, if you’re having trouble getting started with data science and predictive analytics capabilities?
Last few weeks have been very challenging, and at the same time, very interesting! There were several client visits, and in all the visits the key topic was – ‘Analytics as a Service’ (AaaS). It was also interesting to see the analytics maturity levels in these client organizations. Almost all of them have invested in data management platforms, predictive analytics, and have fully embraced both the strategic and tactical value offered. However, what struck me from those client conversations, is their inability to develop business cases and hence, they stayed at the level of discussing a lot about Data Lake, platforms, tools and technologies – strangely the articulation of use cases they want to go after, was never forthcoming!
Hence, I think the idea of ‘Managed Decision Science’ may have some interest to the folks who are struggling to get off the block and start truly delivering business values beyond the rhetorics of Data Lake, Hadoop, Machine Learning, etc.
What is Managed Decision Science?
Just by the first word ‘managed’, you can very well guess that this is Outsourcing – plain and simple. In essence, the philosophy is, find a talented organization (big or small doesn’t matter because we are talking about decision science, and not the number of people they have and how many implementations they have done) that has got the right mix of domain knowledge, data science capabilities, tools expertise and the hunger to work collaboratively to deliver the killer business outcomes.
As a starting point, one can go and look at the Gartner and Forrester recommendations, you will find a list of companies offering analytics services. Well, don’t be amused if you see the majority of the list showing the largest (and most expensive) service providers.
Interestingly, in many of my client conversations I was asked to explain – why many client’s are opting to go with niche & small analytics service providers, as opposed to the big and established names in the services industry!
I had never paid attention to this trend, so a little digging around, I was amazed to see that there are tons of mid-size data science consultancies that offer these services. This includes many who have deep expertise on analytics platform side, some have narrowed yet deep domain expertise in the specific sub-industry analytics use cases, and some have flexible constructs to act as service providers bringing a well orchestrated eco-system into play (connecting the dots between platforms, start-ups, academia, etc).
How do you start the Managed Decision Science Arrangement?
Any company can start an analytics initiative on their own but they face two challenges, both real and perceived.
1.) The real and perceived problem of hiring the right talent in the face of today’s Data Science talent shortage, and
2.) Not knowing exactly how to get started.
A Managed Decision Science arrangement in the form of Build-Operate-Transfer model can run for two or three years, and serve as a quick fix for both these problems. The data scientists can hit the ground running, and start delivering something valuable within the first 90 to 120 days.
The managed decision science partner has something to prove, and ideally they should bring in a team consisting of a set of consulting led/expertise led dimensions: experience in analytics platforms/tools, experience in data science, experience in a specific domain, and experience with the processes one wants to focus on (supply chain, customer experience, price optimization, etc.)
However for the managed decision science program to kick-off efficiently, the program objectives need to be defined upfront with a pipeline of use cases (however crudely defined they may be). For example, some may be focused on supply chain forecasting and inventory turns, and others on new customer acquisition or cross-sell. It’s up to the program owner (which in most cases will be a business champion) to set the agenda with a clear and measurable target of working benefits of the first projects to be demonstrated within every 8 to 10 weeks.
Nice! And seems like all the problems are solved?
Not really. Several concerns I have heard over and over again – how do you identify the use cases? How do you associate a ROI against the use cases?
A critical component of the managed decision science program is the ‘discovery’ phase, wherein a week or 10 day intensive workshop is done with business owners, technology directors and the data scientists, to nail down the initial list of use cases and at a high level, the expected outcomes each use case will deliver. Just to keep the use case pipeline filled up all the time, you may choose to go through the discovery process more than once at periodic intervals.
Also, important is to validate the scope and usefulness of the use cases – if the majority of the use cases fall under the category of descriptive type of analytics – reports, dashboards, and alerts then your team composition for managed decision science will have to drastically change to BI type of skills. On the other hand, if the use cases are of true predictive analytics type leveraging big data, ML and advanced visualization, then you know what kind of skills need to be in the team.
For the great majority of large or mid-size companies that have already started experimenting on decision science type of capabilities, and even the smaller ones that haven’t really gotten off the ground, this sort of BOT model may be just the right approach for lift-off.
Our Managed Decision Science Offering
MOSAIC.DS is our platform-based Analytics-as-a-Service offering. It consists of Decision Science Workbench, Analytics App Market, Gallery of Advanced Data Visualizations, and a Pay-Per-Use Commercial Model.
In the past several weeks, we have started showcasing snippets of MOSAIC.DS capabilities to several clients; the feedback is phenomenal and in general there is an excitement to adopt and solve real business problems.
Wealth advisors are constantly looking for new ways to grow their business, which means not…
What started as health crisis has, in a short span of time, snowballed into a financial turmoil.…
Insurance companies play a crucial role during the current Covid-19 pandemic. They decrease…
In January 2020, reinsurance markets started in an organized enough manner –programs were…