Having experience with different CEM solutions, working both on the CSP side and as a CEM solution distributor, I believe that I gain solid insight into the subject. Depending on the vendor and the defined Use case, it can be implemented quite fast. Es an example, I can share experience of implementing specific CEM solution in segment of mobile data services in less than two months (connecting to the Gn interface of the GGSN), using internal data base and Big Data appliances and without „hardcore "integration with CRM or similar, but rather gathering file dumps on the daily bases. So, in the implementation side, it can be quite efficient. What I found most challenging is CSP understanding of the value that CEM can bring to them and detecting the "quick win" Use case to start with.
I am available to share more info if you would be interested.
For deploying a CEM platform, first the Telco operator must know how to extract all that composes the customer experience data from their network, OSS, BSS, and CRM, and make it seamlessly available in an intuitive IT environment. Most importantly is how to explore the customer's data to get better customer experience view. The problem with all these is the gathering in one place, post-processing, correlating, analyzing, representing and reporting all correlated data. Hence, storage and mining is a big part of any monitoring strategy thus customer experience strategy. Note that the correlation between Network performances and subscriber usage is very important.
Once an operator understands how and where to extract all these data, then a CEM platform/layer can be deployed. In my point of view, this is the fundamental issue to deploying a CEM platform.br,
I can only agree with you, moreover based on observation, I would say customers have a totally different interaction not only compared to last decade but also to last 3 years and, yes, I believe also that in 1-2 years' customer journey will change completely.
In my opinion, CSP's can foresee what those behaviors might be even if behavior is changing rapidly. I will explain below as it relates to your question "what is the point of betting on today's behavior and data processed by old process sequences for survival tomorrow?"
I would answer that CSP's can do a lot by observing past behavior and patterns. However, they must change their approach to data analysis and customer experience and data segmentation. For that, CSP's (especially converged operators) sit on a goldmine composed of:
With a change of the approach in data processing and analysis by correlating the above inputs from all technologies, we could come to an idea of what the behavior/requirements might be in next year(s).
With many different studies done recently, it is apparent that Telco operators are not using the potential of Big Data compared to other industries whereas most of the data sits within them. If we could only change our approach using the data we have, we could deliver much better services/products/offers to our subscribers and make a profound change to our industry or at least give a refreshed look to our industry.
Another point I would stress is that using interviews should not be a tool any longer for CSP's: people are more honest with search machines than with interviewers. Per example, if we are seeking to check user satisfaction for a new connection installed, the operator would get much more inputs by analyzing the subscribers network usage and its interaction with the network (respecting GDPR, understandably) than by requesting a questionnaire to be completed after several weeks of usage.
And yes, given Big Data potential, I believe facts are more important than opinions, it is just a matter of how we process and segment those data.
It is true that, as operators we generally do not exploit properly the wealth of data that we have, and that all those points you mentioned are correct moreover the external influences are very important on forecasting needs, profiling users, and generating insights.1. Outside influencers
I believe the approach to developing a proper prediction analytics (and with it, a customer experience picture and customer expectation) must be changed in the same way Telco Cloud/ NFV is seeking to change the mentality, composition and way that teams are working, and this is challenging on its own. To develop proper analytics we need cross-functional teams, not only teams that work in silos but teams composed of different profiles: engineering (SW developers, radio, core, etc), commercial and marketing, sales. Only when such team work not on project basis but as a fully functional team, we can have proper Prediction Analytics, using AI and ML/DL and generate useful insights, thus improve the customer experience as such.
Now, indeed for a small operator, it might be difficult to create such a team due to its cost.
Definitely this is one of the first problematics but it lies on where the operator/provider stands on its digital path. Namely because of the historical way operators and service providers are organized, today, every silo is an island of its own.
Yes, it is a challenge to normalize data, however wrangling of data is a common process in data science, no matter the format. It just takes some effort and plan ahead; it is not easy but it should be doable. Again a cross-functional team would help.
3. OTTs:Yes, definitely the OTTs are game-changers, but that does not necessarily mean a bad thing. CSP's/operators must find a way to capitalize on the new way user interact with technology. At the end of the day, in order to interact with those OTT's they need to use CSP/operators infrastructure.
Nonetheless, let's start from the beginning: how much does an operator need to know about what subscribers are doing exactly with i.e. Airbnb?
Data that the operator/service provider needs are essentially in its environment, such as:
All that is left is to correlate all these data and the outside factors.
So how much do we need to know what exactly the subscriber is doing on different platforms in order to produce some exploitable insights?
Just to add a note: not only mobile data is important but fix broadband as well for the same persona has a full life: outside using mobile network but also in-house using home network as well (Wi-Fi).
Basically, what I recommend is precisely new approaches, methodologies and tool-sets in order to get better insight and change/diversify/open our industry, but we have to start with what we have and what we have is customer data. This will definitely put new demands and open new opportunities to Telcos.
Yes, routine jobs will be replaced by machines which will then open the doors to the creation of new jobs/profiles and probably some of them will include the management of costumer expectation in a new way.
At the end of the day we are in a transformation phase.
That Netflix looks in the past is understandable, because for prediction you need data that helps you to learn. So why not using historical data if that is relevant to the question to answer.
It will be different when we are using the customer journey method for delivering good customer experience. Hereby, first you define your business story line, then you define the journey for the customer; As third step, define the key data from that journey and measure those data during the moments that the customer steps along their customer journey.This goes along with the other message: "what is the raw customer need"?
With these measured data objects, and I call them the "customer interests", you can only start learning what customers do when they walk through their journey, from this moment on. AI will help you then, only overtime, after you build enough cases that make the prediction statistical significant. You learn about the customer behavior in their moment related to that business story line. There are many story lines, many behaviors, many experiences perceived by the customer. After a while you start learning that some persona like or don't like certain experiences that you, as Telco, deliver. The customer journey tells you then, where the problem is and where the customer_persona get lost or disappears from the journey or accept your story line that you push.
Above is not equal to historical data from the customer; It is more date from "in the moment". Namely, the key data objects define what interest the customer have. And the captured "interest", that shows up during the journey depend on the conversation that you push. So the story line is the starting component and that must fit with the customer intent. This is the area of design thinking what you can use as methodology for developing a fitted business story line. A lot is thinking from the customer point of view, from the external "need" to the internal need (=requests) and map those events at your internal business journey.
For example, the business journey can be a sales, care or a problem solving journey; all with different story lines and all starting with an external customer intent. The experiences you offer the client in your conversation with the client determines if the customer moves with you along your business story. The "intent" is different from a customer request. An intent might result in a request; the request is then a need for information, or a click for putting the product in the basket, or "I agree to buy now". All these steps move the customer with an internal story line along a customer journey that finalize the sales, the care, or the problem solved.