In case you missed it…
So far in this multi-part series we’ve covered:
This is the third and final post in a series on ways CX practitioners can move past focusing on the ‘bigness’ of data and get ahead with smart data techniques.
I’ve saved my personal favorite topic for last: Data integration and linkage analytics. Together, they’re a recipe for smart data that drives positive performance and serves as the most critical step in evolving good service into great service.
What I mean by linkage is simple. It’s the art and science of tying together all existing customer data to better understand the end-to-end customer experience (CX). Compared to the other two smart data tactics I talked about earlier, this one doesn’t have much cost – you just need data and expertise to stitch everything together and analyze it. Easier said than done? Apparently. Even today, very few companies have a systemic process to integrate multiple data streams with customer feedback.
Getting Smarter, Not Bigger
Below are four tactics for adding linkage to your CX and voice of the customer (VOC) programs.
#1: Contextualize: At its simplest, use this easy process split in your roadmap. Upstream effects, which are process data thrown off by CX systems, detail what happens before and during an interaction. Focus here will uncover the why’s of CX and start peeling back root causes. Downstream linkage, how customers impact the business, is the loyalty economics lens where you see what drives financial benefits at the individual and enterprise levels.
#2: Start with the Right Data: To get started, look for these three datasets:
- Financial: Customer and aggregate financial data typically are readily available and help build business cases for CX and downstream analysis. Add a dollar figure to common issues so senior leaders knows where the greatest ROI will be.
- Customer: Often housed in marketing datamarts or enterprise data warehouses for use in other predictive analyses, customer data is relatively easy to acquire and often is structured to match with surveys.
- Process: In particular, operations data. How did a customer navigate to a specific webpage? Which interactions trigger closed-loop? What did staffing look like in the store? What’s the average wait time? Link process to financial and customer data will ensure you’re well on the way to a robust CX linkage structure. You’ll know who was affected, what happened to them, and how it affects the business.
#3: Gain Buy-In: I can’t stress this enough. For linkage to be successful, you must secure senior level buy-in. This step is commonly overlooked.
- Define: Detail what you need to capture, what you’re planning to do, why you’re doing it, how long it will take and what ROI you expect. Be concrete, especially about what the analysis will show.
- Clean and Aggregate: It is an old adage but still true: Data cleansing and aggregation is 80% of the work. Accurately document extracts and ensure you truly understand your data; both will help avoid drawing the wrong conclusions.
- Model: This is where the fun begins. Uncover how VOC is affected by business processes, and how those affect financial outcomes. Avoid over-complicating things. Machine learning may be fun but always go with what’s easiest to understand and consume. Linkage is more powerful when it’s not a black box.
- Act: Analysis without action has no value. What you do with the resulting data – or what you plan to do – is the most important step in this process. Develop a specific action plan with achievable recommendations.
#4: Develop an Ecosystem: Put an ecosystem in place and build it incrementally over time. It should evolve to be systemic and inclusive of both upstream and downstream effects, plus all CX measurement programs. It may take awhile but you’ll eventually have a complete 360-degree view into how a single process change will affect VOC at any touchpoint along the service continuum. You’ll know how that touchpoint will shape CX and loyalty to your organization, and you’ll be able to quantify how a specific process impacts financial outcomes.
Examples
Before I conclude, I want to share a couple of examples. One of our long-term clients had an ongoing issue with customers opting out of the IVR at high rates and defaulting to speak to an agent. When the company analyzed IVR data in isolation, they could see something wasn’t working but couldn’t identify exactly what. They partnered with us to integrate a VOC survey into the IVR; invitations were triggered in part by inactivity and insights shed a little light on why drop-outs were high. Then we combined VOC with process and behavioral data to highlight where people became most frustrated, which modules were most difficult to navigate and which pathways led to higher failure rates. The company then re-designed the IVR over an extended period and greatly increased the amount of people retained, at the same time improving customer satisfaction and saving $20 million in cost avoidance.
The second example involved the head of Care who asked, “How much should I focus on tNPS as a KPI to manage service delivery?” To answer this question, we developed a linkage approach using six months of upstream customer behavior, one-to-one survey feedback on that interaction and six months of downstream data. We were able to demonstrate the dollar value for each point of tNPS; in other words, how customer service affected tNPS and how tNPS then affected the bottom line. Following a poor interaction, Detractors cut spend by four times that of Promoters. We suggested moving Detractors to Passives would be more impactful than moving 9’s to 10’s (aka, extreme Promoters). This is something we have seen very consistently within our VOC practice because not only does linkage quantify the value of tNPS, it also helps guide improvement strategy.
Final Thoughts
I realize this is a lot of information and in the end, what does it all mean? To be sustainable and viable as a field, customer experience has to move beyond simple VOC measurement and real-time portals. Practitioners must incorporate other sets of data into CX. It doesn’t need to be ‘big data,’ but instead ‘smart data’ (aka, using data in the right ways). We all need to ask more “why” and have direct conversations with customers about what they see, experience and feel. Pull in all the unstructured data and quantify it using speech and text. Find root causes and learn where to focus, predict customer behaviors and attitudes. Build linkage into all CX systems. Tie what happens before/during/after an interaction across the customer journey to how it impacts the business financially. When all of these things are a reality for you … it’s really, truly smart data.
I recently conducted a webinar called, “Predictive Analytics: Evolving from Big Data to Smart Data.” If you’d like to watch the recording, click here.
Image source: Thinkstock
