Date Published: August 01, 2012 - Last Updated 5 Years, 190 Days, 5 Hours, 47 Minutes ago
In my previous article, I outlined the many types of data beyond surveys and complaints that should be part of the Voice of the Customer. These included operational data, call monitoring and inspection data, basic contact data (other than complaints), social media and unstructured data and even employee input.
John Goodman at Call Center Demo & Conference 2012:
Pre-Session 7: Occupy the Best Business Case!. A recent study revealed that executives are 2-5 times more successful in getting service issues fixed when the CFO and CMO accept the business case. But most executives have difficulty building a compelling case for investing in great customer service. Securing the attention of upper management – and gaining their buy-in – requires speaking their language and quantifying the impact of their investment. Join John Goodman to uncover the key payoffs of great service for nine different corporate functions. Through case studies, learn how to quantify payoffs such as enhanced margins, reduced legal and regulatory risk, employee turnover, loyalty, word of mouth and social media “connection.” Leave with the tools you need to create a business case that gets results!
The Voice of the Customer must be end-to-end and provide a single picture of experience.
To effectively manage the CE, there is a requirement to have an end-to-end, unified picture of the CE. This VOC allows the organization to:
- Set priorities across all the organizational silos and create the economic imperative for action
- Assign ownership of the quality improvement projects that will enhance CE, and
- Measure the effectiveness and financial impact of the recommended sales, operations and service improvements
The challenge is getting the messy data to fit together, then arriving at a unified decision as to the best opportunities ("best" being defined as doable project in a short timeframe with the potential of significant top line impact), followed by developing and implementing projects to execute on the opportunities.
The term "end-to-end" means from the first pre-sale interaction to the final billing, assuring that all issues have been identified and that the most critical ones are addressed. Sometimes the problems that occur before the sale (which are usually not complained about) do the most damage. For instance, in studies at three different copier companies, I found that sales problems often caused four times as much damage to loyalty, on a per problem basis, as machine breakdowns or repair delays. At the same time, sales problems were seldom articulated as complaints due to fear of alienating the sales rep. Therefore, the most damaging problems are exactly the ones that complaint systems fail to highlight because the customer fails to report it.
A VOC that provides a unified picture of the CE reduced the impact of gaps, contradictions and biases in any single data source. Many companies define their VOC as solely customer surveys, which are a lagging indicator and open to challenge by operational skeptics. Operations data (like warranty claims, missed appointments, invoice adjustments, late charges and missed shipment dates) will flag when a delivery will not be on time, before the problem ever occurs.
The problem that often occurs with multiple data sources is that if they are reported separately, there are contradictions often resulting in paralysis. For example, an EVP of a major communications company lamented, “I get seven reports on quality at the end of each month. Three say 'things are great,' three say 'they are OK,' and one says, 'we’re having a disaster.' Who is telling the truth?" The problem is that, as noted above, when you ask who is in charge of the broadly defined VOC (rather than satisfaction surveys) in most companies the answer is that many functions have a piece of it. How do you reconcile multiple data sources?
Integrate Multiple Data Sources Using the "Multiplier"
The actual integration of the different types of data is also a significant challenge. For example, operations and survey data can usually be projected out to the customer base as a whole. Complaint data, on the other hand, may only represent one out of fifty customers who have had the problem while warranty data usually only represents one out of three customers who encountered a serious failure or one out of ten for minor issues. Unless you have a feel for this ratio (what TARP terms the "Multiplier"1), it is difficult to validly integrate such data with survey and inspection data.
The best approach is to execute a one-time survey to determine the ratio of problems customer encounter to those reported to the company via different channels. Alternatively, you can use a standard multiplier of 10:1 or 50:1, depending on the type of issue. In this survey, you ask a random sample of your entire customer base, if they have had problems, which were most serious and then if they complained. If they did, you ask where (e.g. what touch point) they complained. The result of those questions is a chart like the one below.
The above chart tells you that of 100 consumers who encountered a rude airline gate agent, only 20 complained and that 5 customers went to a social media site, 4 called reservations, 1 went to the frequent flyer 800 number and 1 out of every 500 sent an email to an executive. This 500:1 ratio is what we call the multiplier for executive complaints. Likewise, for complaints to reservations, the multiplier is 25:1, that is there are 25 instances in the market for each complaint received at reservations. We can now use those multipliers to integrate data from all the touch points as well as survey and operational data, as shown in the following chart.
This chart allows us to arrive at the best estimate of the instances of rudeness at this airport by applying the multiplier for each channel to the reports received by that channel. The third column of data above is the estimate of instances in the market based on feedback from each channel. Note that the estimates are not congruent, they vary from 350 to 1,000 instances in the last month – this is a consequence of using multiple touch points.Convert the instance in the market place into a monthly revenue damage estimate. Then, if you know that rudeness causes 20% damage to loyalty and the average airline customer is conservatively worth $3,000 over the next two years, you can now estimate the monthly revenue damage of rudeness at that station. I recommend that you take an average of the six estimates of the instances of rudeness in the marketplace, in this case 555.
555 Customers monthly X 20% decrease in loyalty X $3,000 per customer = $330,000/month.
You have taken the fragmentary contact data; 20 social media mentions, 25 calls to reservations, 2 executive emails and 4 complaints to consumer affairs and translated them into a third of a million dollars per month – enough to get operations’ attention and to allow you to set rational priorities by comparing this damage with the damage caused by other problems.
Once you have the revenue damage of problems, you are now ready to start using the Voice of the Customer to drive the customer experience. This will be the focus of my final article in this series.