MENU

Bringing the human touch to big data analytics

August 18, 2014 • Opinion, Top Stories

Dave in office

Dave Peters, Founder and CEO of Emagine International. (Image Source: Emagine International).

In a crowded market place where differentiation is difficult, average revenue per user is low and price wars are not uncommon, African operators face a tri-fold challenge. They must find new ways to attract new more marginal customers, retain existing customers, and increase revenues. New services and tariffs are one answer, but fundamental to the task is the need to better interact with their customers: to understand them, and deliver tailored services.

Relationships have to be built and protected against both other operators and encroachment from other over the top players. Customer Value Management (CVM), understanding and maximising the value of each individual customer through ongoing, tailored interactions, will therefore be an essential strategy as marketplace complexity continues to grow. It will enable operators to think differently and treat each customer as an individual; marketing to one, not many.

Luckily, technology is now at the point at which it can live up to the promise of this one to one marketing ideal. The effective collection, analysis and application of the treasure trove of customer data that operators have access to is now central to effective CVM strategies. Data streams can be harnessed to create a complete profile of the individual and match marketing offers and communications appropriately. From spend and usage information to location, handset type and rate plan or contextual information such as balance information, location and even the weather, operators are now able to add colour and texture to the profile of the customer and create communication strategies based on what we call the customer’s “Behavioural DNA”.

In its simplest form it might be a notification and upsell message for a customer running out of data, or loaning the customer $1 of credit based on an SMS “IOU” so they can continue to call. At the more complex end of the scale, network dropouts might drive automated credits and apologies to customers with high churn propensity scores.

This has brought the science and speed of Google to the telco enterprise; using machine learning algorithms to optimise the allocation of marketing offers to customers. But as clever as big data analytics and machine learning is, can technology alone address customer relationship problems or opportunities? Sadly not, because as advanced as the machine is, it needs something to work with. While the analytic and machine learning capabilities negate the need for lengthy test and learn marketing cycles, reduce the level of engagement required and make choices more accurate, they need human input in the form of effective, creative messages that are aligned to strategy. And this requires an element of best practice, deep understanding of the customer lifecycle and experience in handling the machine’s output.

As software providers, we therefore need to deliver more than just a “black box” for operators. We need to deliver a “white box” which exposes real learnings together with the service of CVM experts who can assist and educate the operator on what the learnings mean and how they can be best exploited. By integrating a Managed Marketing Operations (MMO) service into the operator’s marketing and IT teams this is easily achievable. Embedded in the operations, they take responsibility for the end-to-end campaign design, configuration, execution and reporting, ensuring maximum ROI on campaigns in the shortest possible timeframe.

Interactions become relevant and valuable and customer experience can be easily tracked and improved, creating rich records of what a consumer wants while creating greater “stickiness”. While this decreases the likelihood of customers churning to a competitor, it can also directly impact the bottom line by delivering a greater return on investment. We find that the combination of onsite managed marketing services together with our contextual marketing and big data analytics platform generate demonstrable incremental revenues for our clients of between three and five per cent each year.

In Africa these marketing skills can be hard to come by, making this type of MMO service indispensable.  For example, we use a combination of experienced Emagine people, with a strong bias towards recruiting and training local talent. Over time this MMO team can be built, trained and eventually transferred back to the operator if they desire.

A good example is Etisalat Nigeria. Initially, the mobile operator was keen to implement a successful set of campaigns to increase ARPU and lower customer churn as part of its growth strategy. Following a best practice gap analysis and implementation of campaigns that focussed on more recharge incentive campaigns, new product and services upsell and additional customer life cycle campaigns, the annual target to stimulate incremental revenue of between two and three per cent of total prepay revenue was met within just six months, as was the target of positive ROI. Revenues are continuing to grow from the campaigns we’re running for them, and our MMO team is still in place, working as an effective and efficient extension of the Etisalat team.

So while we passionately believe that big data and machine learning represent the next step in the evolution of customer marketing for mobile operators; extracting more value from data and enabling faster, more intelligent customer interactions, we also believe that technology is not yet a replacement for the human brain and experience. The two need to work together hand in hand to deliver true customer value and ensure maximum returns on customer value management strategies.

By Dave Peters, Founder and CEO of Emagine International

Related Posts



Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

« »