Predictive Analytics in marketing data lakes and contact centers — My Point of View
I thought of penning down this article after my recent experience in studying the need of interaction analytics in contact centers. There are many research articles, case studies, point of views available on the internet in this domain, but nevertheless wanted to share my opinion.
Firstly, predictive analytics in contact centers should be a high focus area, because this is the place where the customer is directly engaging with you pre/post his purchase. Analytics else where do not have the privilege of this direct contact with the customer. I have been through numerous marketing calls for credit card offers (in an Indian context), which after a point of time becomes overwhelming. Such examples are a case for bad customer experience which can affect the overall perception/brand of your product/service and a poor use case of not utilizing customer analytics.
According to an Opus Research Survey, an overwhelming 72% companies believe that speech analytics can lead to improved customer experience, 68% regard it as a cost saving mechanism, 52% respondents trust that speech analytics deployment can lead to revenue enhancement. Research by DMG Consulting indicates that speech analytics in call centers pays for itself in less than one year, and Techtarget reports that it pays for itself in as little as three months. So it’s no surprise that businesses are adopting speech analytics at a healthy rate. In fact, the market has grown from a mere 24 customers in 2003 to more than 3.5 million in 2015-around 20% of businesses that have contact centers. And adoption is increasing as technology improves, with up to 36% of businesses that do not use speech analytics saying they plan to implement it in the near future.
Customer experience is shaped by proactive and intelligent conversations. The two keywords here are proactive and intelligent. By the time the customer calls an agent, the agent should be equipped with all the necessary details about the customer, his preferences, his purchase history and his contact history. As usual, data becomes a central focus area. Data Lakes needs to be established that would capture and enable a customer 360 degree view from siloed data stores in the enterprise.
Back from my previous experiences with customer data, we know that first party data comes from internal apps in the company where trust is maximum, as they are under the corporate data governance and you can always fix the data. At the same, much interesting things happen in third party data sources, which are ambiguously defined most of the times. This kind of data presents real opportunity, they can be used to enrich your corporate data assets and establish a level of intelligence otherwise not possible solely with your internal apps. Privacy and quality concerns exist in mining these third party data sources such as social media, however the benefits are enormous if you can define a clear data strategy and pipeline processing in combining corporate data assets with third party data. Proactive & intelligent calling is further enabled with insights drawn from the combined data assets.
Lets take an scenario based example, when someone purchases your product on your web store or Amazon, you mostly have the order history & customer information with you. However, how is the customer reacting on your product beyond reviews on Amazon? Can you enable an insight on whether any close friends or family members of this customer brought your product as well? Can you call his friends or family members and market this product? I have not seen an example of such a use case with contact centers, because they normally do not have information on your friends or your family members. Such kind of data exists on third party data sources such as Facebook or in your contacts saved on the cloud. Can this be mined? Of course, you can quickly say that this entails lot of privacy concerns, but nowadays many apps are being given permission to share your contacts anyway. This is just an example, I am not saying that this can be implemented today, but a thought process in this direction can enable more use cases as the next level of understanding the customer.
Social media mining has been talked about a lot in the sphere of digital marketing. Its a modern saying going around that Google or Facebook knows more about you than your parents or your spouse. I have played around the social media feeds, there is some information that can be extracted out of what you are permissible to get from such platforms. Of course, there is a lot of noise as well. The greatest problem is the inherent bias when people post on company official pages, the bias is that they rarely talk or give candid opinions, for example when a company gets an award, all the messages seem to be congratulatory ones. How can any intelligence be extracted from these messages for marketing? apart from knowing the person who posted such a comment is interested in his own company products or services. But lets say when users debate on a trending topic frankly, such as a companies stand towards a political event, the conversations would be long and mature enough for some monetization. So, yes such social media mining involves lot of sentiment and opinion analytics, product review analysis etc.
Of course, there are lot of operational areas like AHT reduction, Xsell etc that large companies have solved and continue to bring improvements. This article is not about what one already achieved, its about what more can one achieve. Such kind of novel marketing comes from understanding the customer, his relationships, his behavior and impulses etc and not just by crude sales and interaction data that typical analysts/apps have in their hands. In such a scenario, one can cross the digital boundary and penetrate into a domain that is traditionally achieved by personal relations like you have with your neighboring grocery vendor.
