#flight-risk

[potatolatte]

If I taught you how to love me, you’d refuse to love me then. So just leave me alone, let the vines grow on my walls. Let cobwebs hang like Christmas wreaths on the walls of my heart. Save your breath and your heart for someone else. I was not made to be wooed, dear boy; I would only chew on your strength. So snap out of your daydream and run away from me, like you would from a hyena wandering down your street.

[israelkloss]

3 Steps to Building Customer Churn-Risk Scores

Follow along with the Scoring Customer Churn Risk deck! 

Did you know that it's possible to put a score on customer churn-risk before they actually leave you (churn)?

 Customers don't tell you when and why they are at risk of switching to a competitor or just dropping out altogether. This kind of business intelligence is becoming a much more important corporate tool.  

You can do this through customer behavioral analysis using tools like Evergage (or even a “roll your own” analytics tool). These tools help you know (preemptively) when a customer's loyalty is at risk. 

Starting out

You will need the following: 

A customer-level behavioral analytics system like Evergage or at the least you should be passing custom variables with customer identifiers into Google Analytics.

At least 1 year of customer-level analytics data from any analytics system. Server logs (such as Apache) are also sufficient.

Historical Customer Data

Historical behavioral data is critical for generating an actionable churn-risk score.  If you don't already have someone responsible for collecting (and protecting) historical data associated with the customers, make it a priority. 

I've written some more about the importance of historical customer data for preventing customer churn here: “Why Sharing (Customer Data) is Caring (about profits).” 

Have a Web Site? You've Got Data!

You might be surprised at how much historical data you're already recording. Got Google Analytics? Got log files? Almost everybody with a website has log files. It may only be Apache log files, but that's still data. You are measuring, whether you know it or not. You've got data. Google Analytics and log files offer more than you may know. 

By exporting and crunching some of the key Google Analytics data I show in this post through a program like SAP or R Studio, you can get more user behavior (and buyer behavior) insights than most people imagine.

Despite this fact, you do need much better than Google Analytics data to stop churn. You really need to analyze powerful analytics that let you segment traffic-to-revenue and advertising ROI data. Google Analytics is just the entryway many small businesses start with before moving on to a more powerful behavioral analytics tool. 

Why Behavioral Analytics Matter

Tracking customer-level behaviors allow companies many advantages: 

1) Preemptive Help

Preemptively help customers (and prospects) by actually intercepting potential problems before the customer is forced to contact you (or worse, leave you). 

2) Customer Acquisition and Retention

Behavior-triggered churn-risk alerts can give Marketing and Sales a heads up on acquisition opportunities and every department in a company a heads up on retention because, of course, customer retention is the job of every department. 

Step 1. Collect and analyze 

If you haven't already, do the following: 

Make a list of your customer's reasons for cancellation 

Make another list of observations from the customer service department (or frontline salespeople) about the frustrations they have heard from customers as with as well their stated (and unstated but apparent) reason(s) for cancellation.

Step 2. Create A Churn-Risk Matrix

Your next goal is to use your existing customer data to put a number on your customer's risk behaviors. 

Below is an example of a churn-risk criteria matrix that we collected from one client's customer service department: 

Diagram 1: Churn-risk Behavior Criteria Matrix 

Take the reasons for leaving that you discovered in step 1 and convert that into an actionable score (from 1-10). Add a heavier weighting to whatever you consider most important, most often observed and most heavily correlated events around customer abandonment of your product or service. 

This may seem more difficult than it actually is. 

For example, let's say that your customer service team (or salespeople)  see a good number of help tickets or phone calls about a certain web site problem. Maybe your records show that there is a correlation between that problem and customer cancellation. 

Note: The way correlation is determined is better suited for another blog post on statistical software packages. But for starters, search for tutorials on R Studio, SAS or SAP. 

Put weighted values on customer events and behaviors that fit your “most important”, “most often observed” and “most heavily correlated” criteria. 

This churn score tool will most likely be used for Customer Service to peg a churn score (revenue critical) to their help tickets (see below). At least that's how we built it to work.

But then other departments will want in on the action, too. So step 3 is a popular next move...  

Step 3. Go Cross-departmental

These scores can easily help multiple departments outside customer service. Go cross-departmental with it! 

Next, we expanded the tool cross-departmentally into dashboards that helped multiple departments get past mere customer monitoring and into full customer lifecycle awareness dashboards (CLAD). See Phase 1 below. 

Phase 1) For Customer Service: A Customer Lifecycle Awareness Dashboard (see diagram below).

Diagram 2: Customer Lifecycle Awareness Dashboard

Phase 2) For Sales: The “Customer Acquisition Dashboard” featuring customer demo data passed into Salesforce from the Evergage bi-directional Salesforce integration.

Phase 3)  For Finance: The “Financial Forecasting Dashboard” (credit: Matt Canterbury). This dashboard uses the customer engagement score from Customer Lifecycle Awareness Dashboard and then applies an actionable probability score that a customer will not pay their bill. 

Phase 4) For Marketing: A “Customer Lifecycle Awareness Dashboard – Social” (see diagram below) to gain greater insight into how their customer's social sentiment and online influence may be improving/threatening new business. (credit: Matt Canterbury and Israel Kloss)

Diagram 3: C.L.A.D. Social 

Wrapping it Up

This was a short look a fun project we did using internal customer data which morphed into 4 cross-departmental dashboards. But each company has different needs and yours may have totally different needs. Don't concentrate on the difference between this project and your needs. Concentrate on using your internal customer data in a deeper way. Customized these 4 use cases for your own business. Build your own internal data-sharing tools cross-departmentally.

Whatever you choose to do, these customer data related issues are not going away for any company:

customer acquisition

customer lifecycle

customer retention

customer churn

social media transparency (sometimes forced transparency)

These issues are all on the rise and are being addressed powerfully with behavioral analytics systems. 

"It is better proactive than reactive. Help customers, preemptively. Shut down data ignorance. Use predictive data measurement to drive your customer satisfaction strategy."

Check out Scoring Customer Churn Risk for more!

Sign up here for the Findable Consulting Newsletter for alerts for the next followup in this multi-part Business Intelligence Series.

[israelkloss]

Why Sharing (Customer Data) is Caring (about profits).

Why Tracking Customer Lifecycle Data = $$$

Tracking your customer's lifecycle means automatic money because it's the key to the Predictive Analytics that help you with "churn prevention" (jargon for preemptively preventing customers from leaving). I'll address more on this in upcoming blogs. Sign up for our newsletter (or to this blog at top) to get notified. 

The Problem

I have a customer with departments that just don't share interdepartmentally about vendors choices and they don't share customer data.

Finance isn't sharing the customer's payment history with Customer Service which, in turn, isn't sharing updates about demo account user's help tickets with Marketing which isn't talking to the Engineering Department which, on principle, never talks to sales (the true front-liners) for input on what the customer specifications that they are actually requesting (don't get me started on that one).

All these departments are heavily siloed (and it's not even at the mid-level company size, yet). And their departmental data is even more siloed. It's a common problem across companies for which I've consulted.

Recognize your department?

Here's how my client's problems play out by department. Do you recognize your department?

Marketing is buying a Marketing Automation tool. But they bought the Marketing Automation tool with the API that nobody knows how to harness. Marketing problem.

Engineering is all excited about their latest SaaS (software as a service) upgrade. But Customer Service can't see the customer's login/interaction records, and thus cannot do predictive/preemptive problem-solving for the customer (before they complain). Nine in ten won't complain; they will just leave. Engineering problem.

Finance just bought a new billing/renewals system but didn't tell anybody about it. And nobody verified that it will work with the Marketing Automation API or with the Customer Service Ticketing System API or Engineering's SaaS database. That's a problem.

These departments aren't working together.

Do you have a Data Strategy?

Do you have a strategy for doing customer data integration inter-departmentally?

Is anyone evaluating the inter-departmental data plan? If not, you're just creating a data-siloed company with no customer intelligence plan and heading toward accelerated churn.

If you have a plan, tell me how it's working for you. Did you find a consultant or software product that really did a good job for you? Let me know!

T-CLAD

Next Up, I'll post some samples from the prototype I built for my customer that gives each department a cross-referenced view into the customer data that matters most to their role.

Teaser: T-CLAD provides Customer Service with a flight-risk score by individual customer (tagged with the customer's risk behaviors) so they can prioritize incoming tickets by customer behavior. 

We call it the "Total Customer Lifecycle Awareness Dashboard" (T-CLAD). Yea, we're still thinking about a better name. :)