Finding Diamonds in the Rough – Are Some of Your Worst Current Customers Your Future Cash Cows?

The oft-neglected customer who represents little value today could be an important asset for your company tomorrow – the trick lies in identifying them.

Companies in the telecommunications and banking sectors that have begun managing customers based on their value have taken that first important step in effectively dedicating their resources proportionately, providing a higher level of service to more valuable customers. This is often reflected through the offering of dedicated customer service representatives, tiered loyalty cards, and personalized products and services, among other practices.

The reason why this is considered only the first step, however, is because one segment of customers is often neglected by companies and not treated in a differentiated manner – this segment, for this article’s purpose, will be called the “future cash cow” segment.

Customers in the future cash cow segment can be identified as follows:

• The monthly or annual value they generate through the products or services they currently purchase does not warrant their being identified as high-value customers.
• They have the potential to:
  • Spend more in the current assessment period and thus could become high-value customers (current share of wallet – short-term potential cash cow).
    • An example of this is a prepaid mobile phone subscriber that splits his or her spending with different telecoms – if he or she chose one provider, their spend would be significant enough to be deemed a high-value customer.
  • May spend more in future assessment periods to become high value customers (future share-of-wallet – longer-term potential cash cow).
    • An example of this is a banking customer that currently generates little value for the bank, but in the future, as his or her wealth and needs change, will generate enough income for the bank to be deemed a high-value customer.

This article will address the identification of short-term potential cash cows, customers which can be targeted through various campaigns and outbound sales efforts to rapidly acquire substantial assets for your company. How longer-term potential cash cows can be identified will be addressed in a future article.

Identification of Short-Term Potential Cash Cows

Among your customers are a significant number of them who don’t give you all their business, for one reason or another. These customers split their category spend in one of two ways. The first, splitting spend vertically inside one product line – an example is the above mentioned case where a customer splits his or her spending with different prepaid mobile phone service providers. The second, splitting spend horizontally across a range of products – an example of this is where a customer holds his or her current account with one bank and his or her mortgage with another.

While there are numerous ways to identify these customers, we will highlight some of the main methods that can be used…

Telecommunications

• The customer generates significant value in one product or service line but does not own other products or services, and, has similar demographic characteristics to those of high value customers. An example of this would be a fixed line subscriber who generates relatively significant value and lives in an apartment building or neighborhood where a majority of his or her peer residents (those of a similar age, gender, household size, etc.) have ADSL packages and mobile phone lines with the company. There is a strong likelihood this customer is splitting his or her spend horizontally among several telecom companies and is a good candidate to pursue as a short-term future cash cow.
• The customer has a disproportionate inbound – outbound rate, receiving a significant level of calls (a level those who are high value customers exhibit) but making few. This may be an indication of the customer using another provider for a majority of outbound calls. Analysis among high value customers to determine a relative corresponding outbound rate to a given inbound rate can help identify such customers.
• The customer has exhibited behavior at any time in the past that corresponded with an annualized behavior which would qualify the customer to be a high-net worth customer. An example of this would be a customer who met the threshold requirements of the telecom operator to be considered a high net worth customer for a short-term period (a month or quarter), but, did not exhibit this behavior over the course of the required period (usually a year). Odds are, this customer has churned some of his or her business to another operator, and, if won back, could join the high value customer segment.

 

Banking

• The customer is receiving a relatively significant salary payment (similar to those deemed as high net worth and only holding a current account with the bank) to his or her current account but transferring it out soon after it arrives every month. These customers are ripe for the picking as their potential value is self-evident.
• The customer is making monthly mortgage payments out of his or her account in installments that can then be used to estimate the value of the mortgage he or she is paying off. Based on the value of the customer’s current account and mortgage, his or her likelihood of being considered a high-net worth customer can be easily calculated. A similar observation can be made about credit card product ownership – electronic fund transfers to accounts in the customer’s name at another bank are a tell-tale sign.
• The customer has characteristics which deem that he or she must be holding other products with the bank, but has chosen not to do so and opted to work with another bank. Customers with similar salary payments, living in a similar neighborhood, of a similar age and household size stand to own a similar set of banking products. Those characteristics common in the various high value customer sub-segments should be identified through analysis. This information can then be used to identify potential high value customers.

 

To effectively be able to do the above, extensive analysis will need to be conducted by a data mining expert. Based on the data on hand regarding the customer, the query criteria will need to be fine-tuned on a case-by-case basis to ensure the most optimal leads are produced, leads which yield the highest possible return.

Generating leads is the first part of the effort in obtaining short-term cash cows – effective sales pitches, product / service bundles, discounted pricing, customized campaigns, etc., will likely need to be utilized to convince the customer to give full share-of-wallet to the company. As it stands, he or she already has a reason for not switching their alliance over to just company.

Understanding the reasons why any given sub-segment of customers has chosen not to give their full share-of-wallet is needed here to ensure the right offer is made to the right customer. Testing offers will be an effective method to determine the right mix of product & service, price, and pitch. Looking at competitor’s offerings is another, particularly around those products or services least held by these short-term potential cash cows.

Utilization of the above listed methods for identifying potential high value customers should become a standard marketing practice in telecoms and banks. Such methods to acquire business and increase the high value customer segment base are extremely cost efficient, and, if tried and tested, very effective.

Ensuring an Effective and Consistent Customer Experience to Realize Improved Retention

Companies need to realize the single greatest asset they have aside from their product or service portfolio are their customer-facing employees, and, accordingly, ensure they are equipped and measured in the best manner possible to satisfy customers to the greatest extent possible.

Think about it – aside from the products or services a company sells, what defines its brand image and perception in the marketplace more than anything else are its customer-facing employees. The depth of a customer’s relationship with a company is to a large degree and extent defined by the experience they have in their interactions with the company’s sales and service representatives. Once a product or service is acquired, these employees are the ones who are responsible for maintaining and enhancing the customer’s relationship with the company.

Yet, time and time again, in every country and sector around the world, it is these same employees that drive customer churn – a recent report estimates that 2/3^rds of customers who sever their relationships with a company do so due to poor customer service[1]. So what’s a company to do?

Get it right through running a project with a three-step approach:

1. Identify existing deficiencies around the sales and service model
2. Make the needed enhancements to the system
3. Ensure consistency through the utilization of effective measurement and rewards mechanisms

 

Identify Existing Deficiencies Around the Sales and Service Model

The first step to take in enhancing the customer experience is to understand the existing pain points – that is, to identify the specific factors causing customer dissatisfaction, or moreover, churn. There are two key tasks here – the first, to identify all customer interactions points, the second, how those interactions are in terms of service quality.

In identifying the pain points, every single customer interaction channel must be examined – i.e. call center, dealer / branch / retail outlet, website, email, etc. – any method in which a customer could experience a letdown and feel the service they are receiving is inferior to what they might find elsewhere or feel undervalued.

To determine what to examine in these channels, the existing level of service must be understood. The most effective way to get this done is to talk to the customer base to understand how they perceive their interactions through these channels. Surveys should be conducted with existing and former customers to understand statistically what exactly is driving their dissatisfaction related to service. This should be complemented with focus groups, aimed at understanding in a qualitative manner what they expect vs. what they received in terms of service. Finally, existing complaints (if captured through recorded calls, for example) should be examined to ensure customer viewpoints are captured to the greatest extent possible.

Separately, management and front-line employees should also be interviewed. They may give insight into why there are customer service-related issues, and could point out root causes which customers may not be able to identify.

Chances are, the findings will show the root cause of customer service dissatisfaction lying in one, if not all, of the following areas:

• Inadequate or Unclear Processes
• Poor Expectations Setting
• Lack of Problem Ownership
• Excessive Processing Delays / Wait Times
• Unfriendly / Indifferent Service
• Giving of Incorrect Information
• Lack of Convenience

 

Make the Needed Enhancements to the System

Once the problem areas have been identified, actions need to be taken to close the existing gaps. Based on the identified root causes, the resolution may lie in improving trainings or findings new ways to motivate employees, in changing certain processes, in modifying specific policies. Regardless of what needs to be done, utilizing best practices will help in ensuring the right actions are taken.

If other companies around the world have been lauded around their excellence in customer service, there’s no need to recreate the wheel in looking for solutions – identifying what they’re doing right around the things you’re doing wrong will help in deciding the course of action to take in determining which changes to make.

More often than not, the root cause of problems will lie within training and motivation. Lack of effective training, particularly around helping employees see how the service they provide to customers affect’s that customer’s perception of the company, likely lies behind much of what is leading to customer dissatisfaction.

Other common problem areas are processes – processes which are not designed with the customer and service in mind, but rather stress security and drive micromanagement. Lack of empowering front-line employees to make a decision on behalf of the customer will definitely come up here – few companies seem to be willing to trust their employees to make a decision without approvals from authorities higher on up.

Ultimately, it is recommended here that changes be made in gradual phases, allowing employees to acclimate themselves with the modifications, and allowing the company to make sure it got things right – a test and learn environment, if you will, should be what’s set up.

 

Ensure Consistency Through the Utilization of Effective Measurement and Rewards Mechanisms

The final phase to enhancing customer experience comes through making sure the changes are effective, are working, and have been accepted across the front-line and by the customers. One method to ensuring consistency in the quality of service delivered across all channels is mystery shopping – a practice which should be business-as-usual in those companies that truly care about the service they provide their customer base.

Mystery shopping (the use of a third party company and its employees who pose as customers to interact with one’s own company’s front line employees to determine the level of service being provided at varying frequencies and in different channels) can also be used to reward excellence in service – recognizing those employees that are truly making a difference in the service they are providing their customers. Random spot rewards (gifts, cash, etc.) should be handed out to such employees, reinforcing the fact the company truly values its employees and their dedication to caring for the customer base.

Other methods of measurement should also become common practice – number of complaints, number of customers elevated, number of appreciation letters, etc. A reduction in complaints will be a tell-tale sign things are headed in the right direction, but should be reinforced with quarterly or bi-annual surveys. For companies in those sectors that can (i.e. banking or telecom), exit interviews should be conducted to continually identify what may still be lacking in terms of service quality.

 

If the above phases are handled effectively, the ultimate measurement will come in terms of seeing a reduction in the company’s churn rate. By getting the basics right around customer service, companies should see an immediate change in their bottom line.

 

---

[1] http://crmweblog.crmmastery.com/2008/11/customer-service-not-price-remains-top-cause-of-customer-churn/

Reselling in Line With the Technology Adoption Curve

Sales and Marketing functions in technology-based product / service companies need to understand their customer base as related to the technology adoption curve and strategically plan their reselling activities around the concept.

For every new product that comes out that can be considered ground-breaking or trend-setting, there is a pace at which customers acquire the given product. This concept, called the “technology adoption curve” is defined as the adoption or acceptance of a new product or innovation, according to the demographic and psychological characteristics of defined adopter groups.

These adopter groups can generally be classified as follows:

Innovators: Wealthier clientele, seeking out and acquiring products as soon as they come out

Early Adopters: Young, well educated and market-savvy, with strong social networks

Early Majority: More conservative, wait for social acceptance of products before jumping on board

Late Majority: Older, less educated, fairly conservative and less socially active

Laggards: Very conservative, technology-shy, traditional

To put this concept into perspective by using an example, when Nintendo Wii was introduced into the marketplace, there was a major rush to purchase it. Those purchasing it on day one were the Innovators, standing in line to be the first to get their hands on the product. The Early Adopters likely purchased it over the first few months of its introduction, quickly sharing their experiences across their vast social networks. The product is likely now in the late stages of adoption by the Early Majority group, influenced by its social acceptance and spurred on by the satisfied Innovators and Early Adopters. In the coming year, the Late Majority will begin acquiring the product, and eventually, few, if any, laggard will join the frenzy.

For every product, the length of this cycle, as well as the percentage of customers who fall into each adopter group, varies. Some products gain immediate widespread acceptance across the marketplace, others take years to.

The purpose of this article is to bring attention to the relation between this adoption curve and reselling. Companies that are able to identify when exactly each of its specific customers purchased a given product or service in the past along the technology adoption curve can then target the same customers with one-to-one offers at the appropriate time along the next product or services’ technology adoption curve.

To be able to effectively resell along the technology adoption curve, a company needs three specific pieces of information around the product / service they are about to release:

1. What the adoption curve for the given product / service will look like (based on similar products or services released in the past as well as sales projections, a company can hypothesize around this, and, accordingly, estimate sales over time periods).
2. Which adoption group its existing customers fall into (again, matching customer purchase timing over the adoption curve of past similar products / services, a company can make assumptions regarding when each of these existing customers may purchase the new product / service).
3. Customer contact information (critical, as no communication can be made, nor past purchases tracked, without it – loyalty programs, warranties, etc. are methods of obtaining such information for retail-based companies – service-based companies should likely have such information on hand).

With analysis around points 1 and 2 completed, a company would then need to devise a contact strategy around its ex-customers, matching the message and offer to the specific adoption group at the right time. The key point is to ensure that the potential for new sales to existing customers is maximized in as rapid and efficient a manner as possible.

A different tactic needs to be followed with each of the adopter groups. Some suggestions:

• Customers in the innovator adopter group should be contacted even before the product or service is released, with offers of trial usage (the first to try it out), pre-ordering services, home delivery, etc. Capturing this trend-setting group’s imagination and interest early-on is the key here.
• With the early adopters group, contact should be made with these customers soon after product / service release. Capturing as large a resell rate as possible to this group is critical – the social networks they influence can greatly impact future sales. To that end, offering discounts, value-added services, and free complementary gifts should be considered.
• The early majority group is a segment of customers which will wait for general social acceptance of the given product / service, and as such, generally do not need to be contacted immediately. However, as this segment is much larger than the innovator or early adopters groups, their importance on the bottom line is significant. Thus, customized offers should be prepared to capture the interest of this group, focusing on the widespread usage and acceptance of the product / service, coupled with some exclusive benefit.
• The late majority and laggards groups come into play usually years after a given product / service’s release. As such, the focus of such an effort should be minimal on these customer segments – predicting their purchase timing is extremely difficult, yielding low response rates.

Companies undertaking such efforts should be careful not to “cannibalize” sales – some early adopters and early adopters will make purchases on their own, with no contact / offer from the company required to trigger the sale. Accordingly, it is important to be careful not to contact customers who the company has already resold to.

Some examples of companies /sectors that stand to benefit the most from such an endeavor:

• Technology Retailers – For example, selling a new version of a given Nokia phone the E71 – to past purchasers of the Nokia E70
• Telecoms – For example, selling a new maximum capacity ADSL bandwidth speed service to those customers already subscribed to the current maximum capacity bandwidth service.
• Banks – For example, selling a WAP-based service it is about to release to customers who are using its online banking services.
• E-service Providers – For example, up-selling to-be released enrollment options offering added value or services to its customers based on their prior enrollment timing and behavior.

Such a concept can be used outside of the technology realm as well, applying to fashion, furniture, airlines, etc. – ultimately, analyzing customer purchase patterns around timing as it relates to any given product or service can greatly help a company in reselling similar new products / services to those same customers.

One final note – the output of such an undertaking should be reflected in the company’s segment-based strategies, in the behavioral dimensions of the customer, alongside his or her value and needs properties. Which segment your customers fall into along the technology adoption curve should be clearly defined, ensuring the information is used on an ongoing basis by segment managers.

Bundling Your Way to Success

Working hand-in–hand, Sales and Marketing teams of companies can generate significant cross-sales / up-selling by bundling products / services together. Of critical importance is piloting the concept to determine the best mix and to ensure revenue optimization.

In almost every sector, there exists a great, relatively untapped opportunity for generating cross-sales. The opportunity – bundling – is the concept of putting complementary products and/or services together and selling them as a package to the customer base. The critical point here is that some type of value is created for the customer, be it a discount, convenience, etc.

Some companies have for years been practicing bundling, and have even been defined by the concept – case in point , Microsoft’s Office software bundle. Their strategy to offer a package of solutions (Word, Excel, Powerpoint, etc.) rather than standalone software has been pointed to as one of the key reasons why companies like Novell or Wordperfect failed with their offerings.

The software sector is only one example – bundling can be seen in practice in other sectors such as retail (i.e. toothbrush packaged with toothpaste), hospitality (i.e. discounted breakfast if packaged with the room reservation), and telecom (fixed-line, mobile line,  internet, and IPTV packaged together), as the concept knows no boundaries. It cuts withing and across sectors as well, as companies can team up to offer bundles (i.e. instant coffee bundled with coffee creamer).

The concept can support numerous strategies, and accordingly, the reasoning for using it should be well defined. The most common reasons include:

• Generating cross-sales: By using the demand for one product to boost the sales of another. An example – bundling a printer with a  laptop in order to boost printer sales, playing to the printing need of laptop purchasers that will develop sooner or later.
• Up-selling: By generating demand for a product through incentivizing the customer base. An example of this – bundling free basketball tickets and free monthly magazines with upscale VIP season tickets for a football club, hoping to drive up demand of more expensive tickets through offering additional benefits in the form of a bundle.
• Product uptake: Bundling a new product with an existing one to generate customer trials and then hopefully future purchases of the product separately. An example of this – a shampoo bundled with a new soap for free, generating trial usage and then uptake in the future.

When designing a bundle, there are two critical pitfalls to be wary of. First and foremost, the bundle should not strip away brand value – customer perception is critical and must be considered when designing the bundle, otherwise, it can harm the overall company value and have lasting consequences. Any cola brand, for example, would have to think twice about bundling its product with an alcohol brand – while there could be a positive increase in sales in the short-term, the long-term harm that could come to the cola brand for being associated with an alcohol brand could be overwhelming.

Second, the bundle should make sense financially – extensive analysis and piloting need to be conducted to ensure this is the case. A certain degree of revenue cannibalization will take place, and could ultimately hurt the bottom line rather than help it. A scenario of such a case can help bring clarity to this point:

In the above case, despite a 1% increase in sales, there is a drop in net profits due to the pricing discount offered to the customer. The reason for this in that cannibalization has occurred – meaning customers who were already going to purchase the two products separately were given a discount that was not needed – they would have bought the products without any incentive anyway. Launching such a bundle would cause detriment to the bottom line rather than helping it, and is a common mistake made by many a company. The best way to avoid committing such a mistake is to conduct extensive piloting to ensure a financially sound bundle has been designed before mass roll out.

In the pilot, the following should be assessed, monitored, and fine-tuned:

• Cannibalization Effect: The correlation between the packages sold and its effect on standalone product purchases, specifically to see the financial impact and the effect on consumer behavior.
• Customer Perception: The perception the bundle is creating with the consumer – is the package “cheapening the brand,” does the created value come across to them?
• Employee Performance: The scripts the employees are using, the messages that work best at selling to customers, the most effective way of convincing customers to make the purchase.
• Packaging: The physical packaging of the bundle, the marketing communications in place to promote the bundle, the location of the bundle.

The importance of conducting thorough piloting cannot be stressed enough around bundling. Only when the value proposition is well designed and the bundle solidly tested and determined to be financially sound should it be rolled out across all sales channels and further supported by more mass-scale marketing communications.

Turning Your Customers (and Non-Sales Employees) into a Sales Force

In a day and age when acquiring a customer is getting more and more difficult due to increased competition and slashed sales budgets, customer referral programs can be utilized to generate significant sales growth in a very practical and effective manner.

Word-of-mouth referrals have been looked at over the years as a key means by which to acquire new customers. The activity of promoting a given product or service to a friend or family member has a significant impact on that person’s likelihood of purchasing said product or service. Inversely, and logically, detraction has the opposite impact, leading to a loss of potential new customers.

A recent study examined the effect of customers who advocate or speak out against companies in the wireless GSM sector in the United States. The findings – customers who promote their wireless company attract a half new customer each on average, and, generate $1700 for the company. Those who speak against their provider cause the company to lose 1.3 potential new customers on average, and generate a net loss of $300 .

Clearly, referrals are an excellent means for generating new business . That said, companies need to proactively push their customers to generate these referrals, but rarely do. Few have successfully tapped into this ever-important lead generation mechanism.

A new and ever-burgeoning means of doing this is through the use of customer referral programs. These programs can be simple or rather complex, rewarding customers with prizes, products, cash, etc., for each referral / lead that turns into a sale. Some examples from different sectors include:

• Bank of America – Rewards both the referrer and referred with up to $50 in cash for the referred opening a new checking account .

• DIRECTV – Rewards both the referrer and referred with a $50 discount on their next bills for the referred subscribing to satellite television services .

• Vonage – Rewards both the referred and referred with two months free Vonage services if the referred subscribes to their fixed line service.

A variation of this program involves opening up the referral program to the company’s own employees, turning all non-sales employees into lead generators. One company that excels at this is PNC Bank. Through their “Chairman’s Challenge” program, the company has brought in over $440 million in new demand deposits and generated $1 billion in deposit and loans balances. The program rewards employees with points for every account that is opened through a referral they submit, then, the points accumulate and can be turned into numerous different prizes .

Companies, particularly those in the service industry, stand to reap significant benefits from setting up a referral program of their own. Some critical principles that need to be adhered to when doing so:

• Simplicity – The program should be easy to understand, easy to use, and easy to process, from both the customer and employee perspective. Over time the program can evolve but should start with a focus on the referral of one product or service, with one prize associated with said referral.

• Financial Feasibility – The rewards provided to the referrer and referred should be financially in line with the profits to be generated from the acquisition. This may be secured through a requirement of some minimal commitment from the new customer (i.e. minimum 6 month contract). Otherwise, in the lack of a contract, the reward can be given out once a certain time requirement has been met by the customer.

• Strategically Sound – The business activity being promoted should be among the company’s strategies, and, should not reward an activity which is likely going to happen anyway. For example, a bank could consider rewarding customers who refer others and set up online banking with their new account. By doing so, the bank would ensure a lower rate of churn (as online banking has a significant positive effect on retention), and would promote an activity they may have a hard time convincing customers of doing.

• Cyclical – Such programs are more effective when they have time periods attached to them, a specific start and end date – thus driving customers and/or employees to take action and make referrals. Setting the program up to be cyclical additionally allows the company to make alterations to the program over time aimed at enhancing its attractiveness.

Certainly, these programs make sense, and will likely evolve and multiply over time across many sectors and countries. Through rewarding its employees, customers, and new customers, it’s the company itself that benefits the most in the end.

Quintupling Your Churn Prediction Performance

As real business cases demonstrate, the performance of predictive data mining models drastically improve when information about historical customer behavior and real-time interactions are put together. Companies with traditional churn prediction models should realize the opportunities they are missing…

What?

Real time business intelligence and data analytics is becoming popular among most leading companies. More and more, focus is being put on the ability to react to customer interactions the moment they occur, rather than later when the real opportunity is lost. Yet, very few blend the benefits of traditional data mining models with real-time analytics. For example, the accuracy of a data mining model that predicts the likelihood of customer churn the next two months can drastically improve when combined with the complaints information during those two months. A red flag should be raised when a customer who is identified to be at risk by a predictive model takes a certain action that is also identified as a customer churn trigger.

But, Why?

In predictive analytics, there is always a balance between how early the model predicts and how recent data it uses as input. The former is a requirement by the business, as it leaves enough time for taking preventive actions. It is also, usually, required as the ETL and scoring processes take too long and can not be repeated frequently over time. The latter, however, defines how accurate a model can be, as the customers usually show their signs of disloyalty when they are just about to leave. The common approach to balancing these two has been coming up with two sets of analysis:

• One for dealing with the real-time churn triggers to handle the recent customer interactions – such as setting warning mechanisms for high-risk complaints.

• Another for predicting customer churn a few weeks – and sometimes months – earlier, using predictive data mining models.

However, these two approaches can be utilized perfectly, to develop data mining models, incorporating the benefits of both. The performance of such mixed models beat both alternatives as standalones, and, in fact, can even multiply their accuracy. A sanitized case demonstrates the impact of mixed-models:

Company X – Middle East – Leading Telecommunications Player

• The company has performed an analysis on the customer complaints and found out that customers making certain complaints have a risk of 7% churn in the same month.

• The company also developed churn prediction models, which identify a relatively high churn risk group, which is likely to churn by 10% after 2 months.

• Putting the churn model and complaints data together for a mixed model yield to identification of the highest churn risk groups, which are likely to churn by 50% after 2 months, if they make certain complaints.

Incorporating the real-time data with longer-term predictive models can save millions of dollars for companies trying to decrease their churn rates, by letting them focus on the best target groups.

So, How?

We recommend three main steps towards building a mixed model for accurate prediction of churn, cross-sales or any business problem of a similar nature:

• Prepare Data: Unlike regular data mining models, the data required for mixed models include both historical and most recent data. While certain customer behavior dimensions should take into account the state of customer a certain time period back (e.g. customer value, product ownership, etc. 2 months ago), the triggers which should be incorporated should be recent (e.g. complaints last week).

• Perform Preliminary Analysis: The preliminary analysis should be conducted to identify the triggers which are relevant for the business problem. For example, if the company wants to predict churn, complaints which are most critical should be identified analyzing their correlations with customer churn. The preliminary analysis should be also performed to understand the correlations between customer behavior dimensions and churn as well (e.g. customers with short tenure are more likely to churn).

• Build the Mixed Model: The approach to building the mixed predictive model, once the data is ready and selected, is not very different from building a regular data mining model. Variables related to churn, from different timeframes, are put together and fed into any of the predictive modeling techniques that are available (e.g. logistic regression, decision trees, neural networks). The result will be models which identify customers at high risk, utilizing both recent and longer-term profiles (e.g. customer who has decreased usage in last 6 months, has short tenure and made a complaint last week regarding prices).

Once the model is ready, it should be put into practice across channels, where they trigger certain business actions, when the customer interaction leads to a significant increase in churn risk, as scored by the model.

What Next?

Gaining ability to accurately identify which customers will churn and taking reactive measures to prevent them is an effective means for customer retention. However, for the best solution, companies should focus on a more pro-active approach and develop solutions for the churn triggers identified during the churn prediction analysis.

To receive more information about our recommendations, and learn about related Forte Consultancy Group service offerings, please contact: info@forteconsultancy.com.

Now, Who Can Sell to This Customer?

Most analytical models developed for customer acquisition, retention or growth do not take into account that it is the human that does the marketing, and miss a great opportunity to boost return on investment. Every call center agent and sales representative is different, as is every customer and without a good matchmaking between them; it is not possible to maximize the conversion rates.

What?

Today, most business intelligence activities in the direct marketing and sales focus on building the ideal list of prospects to sell to, identifying the right channel and offer to use, and in some limited cases, finding the right script to communicate. Although all of these are almost compulsory for effective operations, they leave one very decisive element out: THE HUMAN FACTOR. Ideally, in addition to optimizing all those listed elements, companies should also discover who can sell best to whom and optimize the matchmaking between their sales representatives and call center agents with their prospects.

But, Why?

Due to various reasons, such as demographics, personal history, education and social skills, every salesperson is different from another. Some can better communicate with youth, others with elderly or women, businessmen, expatriates, etc. If half of the sales is about the prospect and the offer to make, the other half is how it is being communicated, which mostly relies on whether the person communicating is equipped with the best skills. Ability to recognize which salesperson is best equipped for which type of customer can lead to substantial improvements in marketing and sales results. If one of the call center agents can relate to and make wonders with the university student customers, why continue randomly assigning middle aged businessmen and retired couples to him or her, when another agent could be performing much better with them?

So, How?

Matchmaking between the marketing, sales teams and the customers follows a similar approach to most optimization problems, with three main steps:

Preparation of Data: Understanding of the performance of sales personnel with different customers requires historical data on personnel’s performance as well as the profile of prospects each personnel has dealt with. Ideally, this would mean availability of campaign management data (who offered what to whom and when) as well as customer segments information based on various dimensions such as demographics, needs, behavior and value. For companies lacking such data today, even collecting it for the next couple of weeks and months with short-term solutions can provide a usable basis for analysis. Yet, ideally, these companies should revisit their data strategies and start systematically collecting these key information elements.

Identification of Factors: Once data is available, the next step is doing a preliminary analysis to understand what factors (i.e. characteristics of prospects such as age, marital status, income level, needs) affect sales personnel’s performance. Performing simple statistical tests or even charting personnel performance across different prospect properties can reveal the most important ones to focus on. For companies with capable resources, building data mining models to identify the factors and segments most correlated with personnel performance would generate better results. Whichever method is used for analysis, a key success factor is the ability to isolate the effect of offer and in some cases the time of offer. An agent could be performing best with the high income prospects, but this could be due to the fact that the agent has been used for communicating an offer only relevant for these prospects lately. To isolate such cases, preferably, all marketing and sales personnel should be evaluated based on the same conditions (e.g. offer, time of day, script).

Optimization of Allocation: The final step is the actual matchmaking, where based on the factors identified and the data prepared, optimal allocation of prospects to representatives or agents is done. Although this is a matter of allocating the best resource for the selected prospect group, it involves simulation and operations research techniques to come up with the best allocation. As the capacity is also a parameter – after all, the number of resources is limited – an agent does not always get the prospects where he/she would perform the best. It is a matter of maximizing the output from overall sales team, not each individual separately. As an example; consider the following scenario, where three agents have different sales conversion rates for three different segments of customers. Ideally, it would be best if both Agent A and C sells to the Youth prospects and B sells to the Middle Aged. However, if each agent can make 100 calls a day and the lists of prospects to sell to include 100 of each segment, this allocation does not work out.

Let’s consider three alternatives for this case:

A. Without Any Optimization: In this case, the prospects would be allocated to the agents randomly, each getting about 33 from all segments. The overall sales conversion rate in this case would be 9.7%.

B. With Best Performer Approach: Since Agent A is the best performer for Youth, these 100 prospects would be assigned to Agent A. Similarly Agent B would get the Middle Aged and Agent C would get the remaining, the Elderly segment. In this case, the overall sales conversion rate would be 13%, a 34% improvement from the random assignment.

C. With Actual Optimization: The optimization results in allocation of Youth to Agent C, Middle Aged to Agent B and Elderly to Agent A. Even though both the Youth and Elderly segments are served by non-top performers in these segments, the overall sales conversion rate in this case would be 14%, a 45% improvement from the random assignment.

As the example above demonstrates, the return from optimal allocation of prospects to sales resources can create substantial impact on conversion rates, which is worth the effort put in for analysis for most companies.

What Next?

We recommend that whatever the size of marketing and sales operations a company has, it should initially perform a basic assessment of performance of its frontline staff across different customer segments. In case significant variance exists across these segments, the next step should be pilot testing the concept to see how much of an improvement it would bring and do a full fledge optimization and roll-out afterwards.

Companies should also leverage findings from these analyses in human resources, recruiting agents and representatives who can sell to the underperforming customer segments or training those who might have the potential to do so. After all, the reason that a company can not sell to certain demographics groups might simply be the fact that none of its sales personnel can click with those demographics.

To learn more about making most out of your call center using analytics, please contact info@forteconsultancy.com.

10 Tips for Establishing a Successful Test & Learn Environment in Retail Industry

Decisions about discount rates, product bundles, offers, changes in the store design, process changes and various other investments are part of daily challenges retailers face with every day. For most companies, unfortunately, the “gut feel” is the main driver for many of these. While the value of experience and expert opinion is unquestionable, sentences starting with the phrase “I think…” increase the chance of failure and leave vast opportunities untapped.

Conducting business experiments are fairly easy in most industries today. Marketing departments have already adapted to experimenting leveraging their below the line capabilities, developing and testing numerous campaigns concurrently in many industries.

In retail, the use of business experiments is almost inevitable, with factors like the impact of the store experience in sales, number and variety of products sold and short product life-cycles hence the need for continuous change. Considering the number of alternative initiatives a retail company can undertake, the only way out is to continuously test, assess and learn.

By introducing the “laboratory stores” concept, some retailers around the world are now testing new offers, products, even change in their HR policies. Leading retailers with different offerings such as Sears, Subways, and Famous Footwear are among those currently using business experiments to test the effect of different business variables on their bottom line figures.

The chairman of the Sear Holding Edward Lampert describes the urge to experiment in stores quite clearly:

“One of the great advantages of having approximately 2300 large –format store at Sear Holding is that we can test concepts in a few stores before undertaking the risk and capital associated with rolling out the concept to a large number of stores or to the entire chain”

In order to develop test and learn capabilities via business experiments, retailers should:

1. Get top and mid-level managers on board, create awareness and demonstrate the value add by business experiments
2. Organize a group of experts who can evaluate and translate statistics into business decisions
3. Establish a platform to track the performance of test and control stores without any bias and to keep repository of the key learnings for the future
4. Standardize the business experiment processes and integrate them into core business processes
5. Share the results, show value add and promote test & learn culture in the company

Here are 10 guidelines for retailers willing to establish a test & learn culture:

Distinguish between what can be tested and what can not

Usually doing business experiments will provide more complete and accurate insights in more tactical decisions rather than high level investment decisions. A well designed test will support a decision on discount rate of an offer but could say relatively limited about a company acquisition, for instance.

Do not try to test hypothesis that are interdependent at the same time

Although there might be quite a number of tests conducted at the same time, interdependencies should be taken into account to really understand the root causes of the business impact. A hypothesis about the shelf location of a product and a hypothesis about the packaging of the same product shouldn’t be tested at the same time since it will not be possible to understand what really caused the final impact; the location of the product or the new package.

Choose test and control stores carefully

Usually companies have tendency to choose the closest or most convenient stores as their laboratory stores. This choice leads to biased results since they usually don’t represent an average store. Decisions should be made by carefully looking at various dimensions like sales figures, customer profile of the store, product range in the store, competitor proximity, employee capabilities and manageability. Try to find representative stores for testing. Once you have identified the test stores, use the most similar ones as control stores.

Effectively track business experiments’ history

Business experiments are meant to learning from. A test result –whether a success or a failure- brings lots of insights for the future decisions. Categorization of those results and storing them in the memory of the company -not the employees’- is crucial for incremental impact of test & learn culture over the years. Make a platform to manage your newly gained knowledge, enabling searching for the definitions, results and key learnings of the previous tests.

Don’t test for the sake of numbers

Although statistical tests and business realities do not necessarily require different designs of experiments, in some cases they do, which requires a fine-balance to get the most out of testing. Testing a new offer on a couple of thousand customers across ten different segments might provide in-depth insights and statistically significant results, but the cost and complexity of such a test may not be justified business wise. On the contrary a handful of observations will not be statistically significant although it might be adequate to make an accurate business decision in some cases.

Standardize the test process and embed it into the business as usual

Create standardized processes regarding the business experiments like selecting the test sample, deciding on measures, and analysis of the results and embed them into the ongoing business processes. Processes like campaign launch, offer development or commission structure changes should refer to test & learn processes at some points. Managers should demand these analyses during their decision making and approvals should be linked to the test results.

Automate for the test & learn culture

Usually one of the greatest challenges while establishing the test & learn culture is the load of manual operations that is required to get a clear test case. For the companies which are commitment to this culture, the ideal solution is the automation of measurement and analysis as part of their reporting environments. This will not only make it sustainable but will also prevent rework and possible data quality issues. Integration of the POS systems with test and learn platform is a good start.

Make sure that the test & learn process is user friendly

Although test and learn process includes statistical or analytical elements, it’s the business departments who really need those tests and will use the results for the companies’ benefit. Translating the language of statistics into business words will remove the barriers and will get more people on board for the business experiments.

Don’t suffer from paralysis by analysis

Speed is one of the must-haves in the retail industry. Companies should make sure that they are not loosing their competitiveness in speed of decisions, while doing lots of tests and analysis. Getting paralyzed with more than enough tests is as bad as following only the guts feel.

Train your employees on basics

Like many things in retail environment, test & learn culture is very dependent on the people. At the very basic level, all relevant parties should be speaking the same language when it comes to business experiments. Even an employee who has no responsibilities regarding business experiments should be aware of terms like hypothesis, control store, sample size or lift of the test. What people don’t feel comfortable with, they become either afraid or ignorant of.

Get a Free List of Your Competitors’ Best Customers, Today!

Many companies already own the right data for targeted acquisition from their competitors, yet most aren’t aware of it. Is your company one of them? What you think you don’t know but actually likely do regarding your competitors’ customers represents a huge untapped potential that could create substantial impact to your company’s bottom-line.

What?

A number of industries, especially telecommunications and finance, are facilitators of interactions between people – be it, for example, a phone conversation, or, a financial transfer.. These types of interactions allow such industries to have a unique ability in terms of marketing: direct access to competitors’ customers. When a telecom operator’s customer makes an off-net call, or when a bank’s customer makes a money transfer to another bank, they provide precious bits of information for the company – the phone or account number of a potential customer as well as behavioral information about that potential customer. Using a blend of traditional and unconventional tools of data mining and direct marketing, it’s possible to reach out to these potential customers and make very specific and targeted offers to them.

But, Why?

The utilization of analytics in designing and conducting marketing activities has become a de-facto standard among the best of the best, providing significant benefits to those organizations wise enough to realize its potential. Mainly until now though, most of the analytics-driven marketing activities have focused on the existing customer base – for retention, for internal growth, and sometimes, for win-back. Many of the companies in the aforementioned industries have thus far wasted the opportunity of using analytics for acquisition. If data mining techniques have been useful for identifying untapped potential in one’s own customer base, why not use them to get a better understanding and targeting of the competitors’ customers interacting with one’s own?

Using already accessible internal data to cherry-pick the competitors’ customers provides a highly cost-effective means for acquisition. It also allows companies to select targets for acquisition that are most related to its own customer base, hence increasing the loyalty of its existing customers through the building of a closer-knit community.

So, How?

Similar to most customer analytics initiatives, competitor customer acquisition starts with preparing the data required for analysis and targeting. A competitor customer data mart – a data set including one potential customer on each row as well as summary of his/her interactions with your customers – is best-suited for this job. In this competitor customer data mart, you would have:

Telecommunications (from CDR data)

• A unique identifier: Phone Number of the Competitor Customer
  
• History: A field regarding the length of time in years the phone number has been appearing on your network as a called individual.
  
• Value determinants: Fields regarding count, duration and value of interactions with this customer from your network (e.g. different number of your customers calling the number / total MoU for calls to the number)
  
• Behavior determinants: Fields regarding time and type of interactions with this customer from your network (e.g. SMS interactions mostly / weekend-heavy users)

Finance (from Transactions data)

• A unique identifier: Account Number of the Competitor Customer
• History: A field regarding the length of time in years the account number has been involved in financial transactions with your customers.
• Value determinants: Fields regarding count and monetary value of interactions with this customer from your customer base (e.g. different number of your customers transferring / total $ of transactions with this account)
• Behavior determinants: Fields regarding nature and type of interactions with this customer from your customer base (e.g. small and frequent quantities / currency used in interactions)

Once such a data mart is ready, the next step involves the use of traditional analysis and data mining techniques – such as value and behavior based segmentation – to identify the best targets for acquisition (in addition to business case modeling to understand the potential revenues and impact on cost of acquiring a given customer)Usually, the competitors’ customers with the highest amount of interactions with your customer base would turn out to be the most valuable customers of your competitors, hence the best targets for your acquisition purposes. Other factors of course need to be examined (i.e. the benefits of not paying an interconnection fee in telco, for example). Based on the behavior segments in your target base, you can approach them with value offerings that are most relevant for their needs (e.g. offering weekend discounts to potential customers who interact with your base most frequently during weekends).

Of course, the natural question at this stage would be: “Now that we know whom to target and what to offer, how can we communicate with them?” Two alternative answers exist for this question:

1. In countries where rules and regulations allow such actions and the local culture is such that the potential customers would not be irritated, the most effective approach would be to reach out directly. In telecommunications, this means calling them or sending an SMS to their phone numbers – which is already known in CDR data. In finance, this would mean either making use of contact details provided by your own customers when performing their transactions, or making dummy transfers towards your potential customers – such as a $0.0001 money transfer to their account with a personalized message and offer as the description of the transaction.
2. When existing regulations or local culture does not allow for direct communications with your potential customers, the next best alternative is using your own customer base for contact, through the leveraging of referral programs. Once you know which potential customers you desire, it’s easy to identify which of your own customers interact with them the most. Using highly targeted referral offers – such as ‘get the last customer you’ve called on to our network and you both get 200 free minutes’ – your customers would literally work as your intelligent acquisition channel, grabbing the most valuable customers from your competitors.

What Next?

Using internal data for competitor customer acquisition may seem to be an unorthodox method for most traditional marketers. Yet, as long as regulations allow for it and you avoid invading the privacy of customers, it can generate quick profits and build an avalanche impact, as the more customers you get, the more visibility you will have over your competitors’ base through their interactions. If you are up for it, we recommend that you start with some quick-wins and test the concept in your market.

To learn more about acquiring from your competitors using analytics, please contact info@forteconsultancy.com.

About Forte Consultancy Group

Forte Consultancy Group delivers fact-based solutions, balancing short and long term impact as well as benefits for stakeholders. Forte Consultancy Group provides a variety of service offerings for numerous sectors, approached in three general phases – intelligence, design, and implementation.

The Single Easiest Way to Grow – Winning Back Lost Customers

Companies often fail to tap into and benefit from the data they possess. Possibly the single most ignored information pool is that about a company’s ex-customers. Such data, if used properly, could lead to the acquisition of a significant number of customers and the generation of significant revenues.

What?

• Companies typically lose 50% of their customers every 5 years
• Companies have a 5 – 20% chance of turning a prospect into a customer
• Companies have a 20 – 40% chance of winning back an ex-customer

The facts speak for themselves. What’s baffling is that most companies do not actively address these facts. Sure, many companies have some formal type of retention effort in place, trying to actively reduce churn, but how many actively try to win back customers? It is a company few and far between that has employees dedicated solely to winning back customers. In light of the above facts, one wonders why.

But, Why?

It’s relatively easy, and, relatively cost-effective. Whereas acquiring new customers involves significant media spend to get their attention, winning back lost customers could be as easy and cost-effective as making a few phone calls. Further (depending on the sector in question), new customers need to be on-boarded, with numerous additional interactions to sign contracts, inform them about products and policies, etc. Lost customers that return, on the other hand, can usually be brought up to speed in a much quicker manner due to their familiarity with the company and its offerings and methods.

So, How?

Winning back customers is essentially an art that needs to be fine-tuned through a trial and error effort. At play are three factors that need to be tested:

• Audience – Not all customers can be won back, or even should want to be won back. In defining the customer segment to tackle through winback efforts, the lost customer’s information completeness (their contact information, their product / service preferences, their reason for leaving in the past) and past / potential profitability need to be considered.

Those customers who cost more to service than the revenues they generate should automatically be excluded from the winback effort, as should those customers you have little to no information about. Logic dictates the more information the better, as when the pitch is made it can be tailored to that specific lost customer’s preferences and reasons for leaving. Trial and error efforts will ultimately determine what information is a must have in trying to win back customers, with the winback ratio (% of customers pitched to who accept the offer) then deciding the specific segment(s) of customers to focus on.

• Contact Method – Customers will have a different reaction and acceptance rate based on the channel used to make the winback pitch. Traditionally four methods have been used to try to woo lost customers – face-to-face, phone, mail, and email. In a given company’s specific situation, lack of customer contact information may dictate the method used. Depending on the pitch itself and the audience, different success rates will be realized through different channels. It is important to test the numerous variations possible to ensure the most effective (cost and winback ratio-wise) method is used before ramping up the winback effort from a pilot to a full-scale engagement.

Those companies lacking information about their lost customers must immediately begin addressing this fact. With no way to get in touch with the lost customer, a winback effort is almost impossible to conduct. In sectors like telco, finance, and travel, lost customer information is traditionally on file (or should be captured when the customer is churning – at a minimum, contact info like mobile phone number and email address). Sectors like retail and energy rely on loyalty programs to not only capture customer information but to also track their purchasing patterns.

• Offer – Often the most difficult part of the winback effort to design and get right, the offer plays possibly the biggest role in determining your success rate. It needs to both address the reason why the customer left in the first place as well as offer them something not traditionally given to any average existing customer. For example, if the reason the customer churned was due to poor customer service, the pitch could possibly include some type of apology as well as the offering of a highly trained and dedicated customer service representative. If it was due to price, on the other hand, it could include some type of discount for a given amount of time.

When the churn reason is unknown, analyzing the customer’s product / service usage patterns is one way to customize the offer. For example, a customer who churned from a bank who conducted money transfers at least 4 times a month can be won back with an offer to give up to 5 free money transfers a month for a year.

The more fine-tuned the offer is to the needs of the lost customer the higher likelihood he or she will accept it. Again, a trial and error effort will go a long ways to determining the right offer to each customer segment.

It is worth noting here that a lost customer should be viewed as foregone revenues – any type of profit generated from such a customer is acceptable. Not to say that the offer should be such that the company barely breaks even with the won back customer, but giving back generously to lost customers will ultimately be worth it.

Other factors to consider include timing and frequency. Depending on how long it has been since a customer churned, different results may be realized in the winback effort. Based on the trials, the ideal length of time following the churn event to reach out to the customer with a winback offer should be identified (i.e. 3 months after churning a winback offer should be presented to the customer). In regards to frequency, the customer may reject the initial offer – a follow up a few months later should be considered. However, there’s no sense in beating a dead horse, and thus, if a customer rejects two to three offers to return, he or she should be considered as “unwinnable,” with no further effort spent pursuing the winback.

Finally, it is critical to incorporate winback efforts into a “business as usual” model, with targets set and premiums offered to the various agents who will contact the lost customers. Just as there are targets around acquisition and retention, so should there be around winback efforts. The more the activities are considered and treated like a part of everyday business, the more successful the efforts will be.

What Next?

Using existing data about lost customers, a company can quickly increase their revenues as well as their overall customer base. Marketing and Sales need to work hand-in-hand to tackle this worthwhile effort, as both have an important view on the issue – Marketing the expert on what the customer would like and on conducting trials, Sales on how to pitch it. Such an effort can be designed and rolled into a test phase within weeks, and can pay for itself within days.

To learn more about winning back lost customers, please contact info@forteconsultancy.com.

About Forte Consultancy Group

Forte Consultancy Group delivers fact-based solutions, balancing short and long term impact as well as benefits for stakeholders. Forte Consultancy Group provides a variety of service offerings for numerous sectors, approached in three general phases – intelligence, design, and implementation.