New opportunities, like rapidly expanding sources of customer data, bring new challenges; there will be new requirements, with new skills and new experience required to handle this. In some cases you may know what needs to be done, and the output required; in other cases you may not. Sometimes you may personally know where you want to go, and you may have the hard skills to do it; but you may not have the team to support you.
We are seeing more and more companies looking to ‘big data’ for answers to better understanding customer behaviour. When this new ‘big data’ opportunity is identified, there are various questions that often get asked:
- Do I know what I want to get out of the data
- Do I have access to the right resource
- Do we have access to the data itself
- How much value will this all add
In a recent example of this, a financial services provider decided they needed to show their insurance clients that they could add value to their data. They had a team of actuaries, and much data on their large policy base. However after three months of applying multiple advanced modelling methodologies they found little more than truisms like:
- Older individuals were more likely to become ill
- Lower affluence individuals lived in lower affluence areas
The financial services provider’s response was ‘go a level deeper’, but the actuaries did not have the technical skills, or the data understanding, to do this. Moreover they did not have the vision of how to create that elusive value.
This is a common trap that comes from not asking the right questions at the start. A better approach could be to ask ‘what would change the way we make decisions about individual customers today’?
In some cases this may be obvious; in our financial services example with thousands of insurance policy holders, there can be millions of pounds at stake. A small tweak to retention rates, better conversion of new customers, better risk modelling, can each lead to huge benefits.
For other industries it may be a little more complicated. For instance a hotel chain’s longer term objective may be to ‘increase their share price’; to do this they may need to increase the value of the hotel portfolio, which would in turn be done by increasing the occupancy rates and REVPAR (revenue per room) at each hotel. To do this they need to understand what customer segments are adding the most value, and to set about recruiting more of these. To understand the segments they may need to import external life-stage and affluence data, and match it against their existing customers.
What does this hotel solution then look like?
So our new more focused goal is to increase the value and volume of sales from those segments that yield greater value. This could also be broken done by:
- Getting more sales from existing customers with potentially higher value
- Recruiting more customers from these segments
How would this approach be applied in practice?
Any insight solution provider (internal or external) needs to be given a clear (or SMART) objective, and they need to provide a clear step by step way of getting there.
Remember there always need to be an element of practicality!
If you can truly picture both the inputs, the outputs, and the detail of the process in-between, then you are ahead of the game. In many examples of new projects, this thinking is found to be lacking.
So before you start thinking of specifying infrastructure or tools, you need to consider the people required, and their skill sets.
Let’s consider the skills requirements for the hotel project:
- To define a set of feasible outputs that would add value
- To identify a process for doing this
- To review the inputs, quality check them, and identify any other required inputs
- To prove the value of the new insight development process, which may include
- Running a POC in a static environment
- Manipulating large volumes of data
- Sharing results with internal and external stakeholders
- Delivering to the brief
- To standardise all this, so that it is repeatable
- To turn it into a BAU process
The industry term for someone who contains all these skills is a ‘Unicorn’, as they are hard to find. However in my experience it is not that these people don’t exist; they have often been promoted to a point where they have lost touch with the detailed nuts and bolts of the insight projects they are managing.
Building your ‘big data’ team
One option is to employ a whole team for this new area; however not only is the recruitment process timely and costly, but you will need to find the right leader, and there is always a risk that it will cost a great deal of money and not deliver what you need.
Alternatively you could consider kick starting things with contractors, but by the point you come to get your permanent team on board, you will have had to made decisions on infrastructure, and developed process, that the permanent team member may not like. Besides when the contractors leave, much of the knowledge goes with them.
Using an external consultancy can get around many of these issues; they can be as flexible as contractors, support ongoing BAU, and reduce the risk of starting the new team. However they are expensive (there is a reason that dunnhumby sold for £2bn); so you need to ensure that you are getting the value you need to justify the investment, and to make sure that the IP is kept inside your company.
Looking at service providers for other sectors, there could be a better way.
Oliver (http://www.oliver.agency/en) is a new type of consultancy/agency designed for those companies who want to take their marketing in house, but with external support. In other words to build an in-house agency.
Oliver will put their own team leader into an organisation, and they have the necessary skills within Oliver to run all the campaigns an agency would, enabling you to get started quickly. However over time they select the right team members for you (to be employed by you), and slowly the team morphs into a mixed team of external consultants and internal FTE’s.
This approach seems to be the best of both worlds, quicker and lower risk set up, you have access to all the skills you would need from day one, you retain the knowledge (unlike contractors); and it is scalable as you can recruit internally.