One of the biggest challenges CX practitioners face is uncovering actionable data that will help improve the customer experience. Often company structure and culture play a big role in determining how hard/easy it is to dig out the data. In this article I’ll help you understand where and how to find the data, and how to organize it to become useful to improving the customer experience.
CX (customer experience) and CEM (customer experience management) practitioners require actionable data to understand, interact with, and plan to change the experience of the customer. The customer journey is full of interaction points between the brand/company and the customer. So where do CX and CEM professionals get actionable data to make intelligent decisions?
After looking at business data for more than 25 years I find that most amalgamated companies have the best opportunity to gather data freely, while companies that have intentional divisions with P/L responsibilities most often have siloed data — then there is the occasional start-up where the data was amalgamated from the beginning. This creates 2 different environments for CX and CEM professionals to practice in and two strategies for getting actionable data. In the amalgamated company data is yours for the asking in the divisional company data must be tracked down and sometimes permission must be obtained to use it.
So where are the data sources?
If we get or hunt for data and organize it by typical customer lifecycle journey phases we will be miles ahead toward the finish line and start with a good model for developing CX and CEM data and identifying the sources. Data comes from different sources depending on the phase of the lifecycle:
- Awareness and consideration are increasingly becoming digital and there are many tracking software packages that collect digital click streams on a 1:1 basis.
- At first purchase and repurchase there is point of sale software packages.
- After first purchase it becomes harder and more creative to track usage and create attribution through win-back and repurchase.
Then the practitioners are confronted with the issue of connecting the dots from awareness through repurchase from the many data sources.
A persistent customer ID is required to connect the dots from awareness through repurchase; a unique identifier that allows data from disparate systems to match the activities of a single customer. Now that we know we need to track a customer through all lifecycle phases with a persistent ID we have the problem of grouping customers in large enough units to be able to focus on them.
Of course you are thinking “why can’t I just manage the single customer and forget about grouping?” Managing customers 1:1 is a matter of logistics and energy. I think we may agree that it might be ideal for the customer to journey through such a personalized environment, but 1:1 is not manageable for 99% of businesses. The most businesses seem to do is mass customizations based on some groups of attributes. The reality is CX and CEM professionals cannot cope with 1:1 data because there are too many details to manage and too many decisions to make. Yes there is software to deal with a click stream like environment and we are seeing artificial intelligence to learn about the individual – but the world is just not there yet — the reality for most businesses is we need groups of customers to think about.
Data comes from many places as described below and those places tend to be specific to lifecycle – the sources listed below are mostly for CEM especially between use and win-back.
Does Big Data or Click Stream Data “come to the rescue?”
There has been a great deal of buzz in past years about big data and the ideas around it. The promise of big data for CX and CEM is that knowing as much as we can about “the customer” leads to a better relationship and/or more sales. Creative ways of describing big data such as “gardens, lakes or reservoirs of data” tries to make big data sound more desirable but big data is not formatted in any way that is actionable and consumable by CX or CEM professionals. If data needs to be summarized, rolling it up to consumable groups of people with testable attributes, lakes and gardens don’t help CX unless they transform data into actionable summarized small data blocks of interactions and attributes.
“Ultimately CX and CEM professional needs logical groups of people as data with sufficient attribution to make manageable decisions – you may find big data and click stream data muddy the water is you have developed segments but don’t track by segment.”
Typically, personas are considered aspirational and segments attributed by CX/CEM professionals. Those practitioners on the CX side tend toward persona and consumer research like data and those on the CEM side favored attribution like data from learning what the customer has done and where the customer has been.
The level of data “roll-up” required to create a persona or segment may be in the tens or hundreds of thousands of records and sometimes even millions of records. CX and CEM data does not need to be rolled up to personas or segments; but at least rolled-up to logical grouping of attributes. As an example: if 5% of customers take a certain path in a digital experience that leads to a conversion, that group can be considered large enough to create programs for. In theory this gives the CX professionals < 20 personas/segments to deal with. 3-10 personas, segments or logical groupings are desirable to make creating programs manageable.
CX professionals will be looking at market research, looking forward to customer activities while CEM professionals will be looking at testing/learning data and looking back at whether a customer was successful or not – both professions plan for the ideal experience.
Many software companies specialize in artificial intelligence (AI) watching digital click streams such as a customer moving through an ecommerce website. That practice is often 1:1 between the AI and the customer. Those companies are able to change the experience (intervene) based on decision models. That said, CEM professionals most often work at a much higher level than 1:1 data can provide. 1:1 Data must be rolled up to a logical group to become consumable.
I want to call out one industry that seems to never want to get beyond the intervention point – “Retail” especially legacy brick and mortar retail who seem only concerned with “more sales.” The reason I call that out is if all you want is more sales then doing CX/CEM and all it promises is tainted by greed and has little to offer the customer; just having an AI apply decisions sounds like something out of Frank Herbert’s novel “Dune” which needs no people at all on the CX/CEM side.
There remains an “art” to actionable data not yet described by AI, knowing which data to apply is a process of learning and relationships are complicated so human understanding and emotions play a role. The human mind decision making process passes through the emotional center first before logic. To serve the human customer “big data” needs to become “rolled-up data” and “small data” need to become “grouped data” to make all data consumable to CX and CEM practitioners. Given the reality of business today all that data must be skillfully and artfully applied with a human touch. CX tends to be on the brand and emotional side while CEM tends to be more data centric.
Data needs context in the CX/CEM practice – below is a chart describing the development of personas and segments in the context of customer journey lifecycle. Records are aggregated in tens of thousands and millions to get consumable and actionable data for logical groups. Remember this is at least a statically accurate sample of data, not necessarily all data available; but it could be all data available too in smaller companies.
Where is the data?
How do you get data and what data do you need to be successful?
Given that I can describe an interaction between a company and a customer with just 2 data points what is all the fuss about data? All I need is the name/title of the interaction point and a reference to a logical group. I wish it were that simple; the reality is “just knowing about an interaction or activity does not make it actionable.” CX and CEM professionals need sufficient data to make decisions.
Basic data for each interaction
- Describe the interaction point
- Describe the logical group(s) – persona(s)/segment(s)
- Identify barriers and pain points
- Identify moments that matter and accelerators
- Identify opportunities to gather attribution and listen to the VOC
- Did the interaction lead immediately to or ultimately lead to a conversion?
- If ultimately what are the steps in the chain of interactions?
- If so how much revenue did the conversion generate?
- Is this interaction point planned or unplanned?
- If planned what was the program that created the interaction
- Is the interaction the next step in a chain of events
- How much did the interaction cost the company?
- Is this interaction inbound, outbound or interactive?
- How often does this interaction occur?
- What was the emotional state after the interaction?
- Give a score(s) to the value of the interaction from the customers standpoint
- Give a score(s) to the value of the interaction from the company standpoint
- Give context to the interaction by the lifecycle journey phase
- Give context to the interaction by marketing channel or media tactics
Non-typical basic data that is easy to imply
- What was the logical group(s) doing, thinking, and feeling
- What were the logical group’s objectives and needs
- What company assets or technologies were needed to support this interaction
- Who in the company where customer facing during this interaction
- What is the emotional state before the interaction?
Typical non data elements
- Create a recommendation for change
After the fact data collection
- VOC customer survey data
Tracking a single customer then rolling up that customer to a segment (a logical group); takes some study with test and learn discipline to derive the “right data.” Digital customer experiences offer a simplified path to gathering data while analog experiences need to go a different direction. One way or the other you must define useful groups. In the CX environment the groups are aspirational – you get to make the groups up then apply research to prove your thesis. In CEM the path is through learn and test data which grouped logically. These lists are by no means “comprehensive.”
- CX data sources actively used
- Consumer research
- Instore cameras – path analysis
- Focus groups
- Buying behaviors
- Usage behaviors
- Value needs and behavior studies
- CEM data sources typically used
- Google Analytics
- Mystery Shopping
- VOC/I/E Survey
- Store Path Tracking
- Click stream analysis
- A/B Testing
- Shopping Cart
- Social Sentiment
- Post purchase Survey
- Mystery Shopping
The diagram below describes many data features of CX (above the dashed line) and CEM (below the dashed line) that lead to the star of the show “the interaction point” or IPoint. The IPoint is where the customer and the company interact. Each IPoint has a context within the lifecycle. I realize this is an eyechart but the reality is there are many activities and processes that can provide some actionable data and at some point the lines are blurred between CX and CEM.
Someday AI will be smart enough and big data useful enough to make an impact on the positive side of the customer experience so I find experimenting with big data and AI useful to these ends. Right now companies need data which is sometimes siloed – that data is a gift to the company CX and CEM practitioners who can skillfully apply what data you find on behalf of the customer experience. We showed you where to find it, how to use it in its basic application, and how to create useful models for interacting with your customers.
Share this Post