Stopping Churn in Its Tracks – Proactive Retention Strategies for Mobile Operators

While almost every mobile operator has a reactive retention strategy in place for preventing subscriber churn, few have developed proactive retention strategies for doing so, failing to benefit from this practice proven to stem the tide of customer loss…

Retaining customers in the telecom industry is becoming increasingly difficult. Not only are competitors able to mimic tariffs and offerings at the drop of a dime and thus diluting product / service differentiation, subscribers are only a phone call away from porting out to them.

The process for switching between mobile operators is now more easier than ever – in Canada, the mobile number portability (MNP) process only takes 10 – 20 minutes to complete, with the subscriber ported over and using the services of their new subscriber in less than a half hour! While MNP in most markets won’t become this easy for quite a while, it’s quite clear that the churn issue will continue to grow for mobile operators around the world.

For many years now, mobile operators have been trying to address the churn issue through reactive churn efforts – that is, trying to convince a subscriber to not close their line when and only when that subscriber calls the contact center or visits a dealer to do so. The traditional method used to stem churn has been to offer the subscriber some incentive to stay (like a handset if he or she signs a year-long contract, or, a certain number of minutes for a certain number of months for free).

While such efforts have been effective in some markets at improving retention rates, they are not, and never will be, in others. The reason is quite simple – reactive retention strategies only work in trying to keep postpaid subscribers. The strategy requires the operator have a chance to make an offer to the subscriber, a chance which presents itself when postpaid clients give notice as to their desire to churn. Prepaid subscribers, on the other hand, just churn – they simply put aside their existing SIM card and start using a new one from a competitor. And in most markets globally, prepaid subscribers dominate the market – in Italy, Mexico, and India, for instance, more than 90% of the market is prepaid.

As such, mobile operators (with the exception of those in markets like the USA and South Korea, where most subscribers are postpaid) are unable to intervene in preventing a large majority of their subscribers from churning at will. What’s a telco to do? Launch proactive retention strategies.

Proactive retention is all about reaching out to a subscriber and addressing churn before it becomes something a mobile operator has to react to, thus the name. Rather than wait for a subscriber to…

  • Port out, using the process defined by the given nation’s telecom authority
  • Call the contact center or visit a dealer to close their postpaid line
  • Pull out the SIM card in the case of a prepaid line

…a mobile operator can proactively contact those subscribers it predicts will churn in the coming weeks or months and try to stop it from happening.

To set up proactive retention strategies, we recommend mobile operators follow the below four steps:

1. Effectively Define Churn

As the first step in building a proactive retention practice, mobile operators need to first define properly what churn really means, for both prepaid and postpaid subscribers. This is not as easy as it may sound, as the definition used by telecoms globally can be quite different. If a prepaid subscriber does not make a chargeable transaction (i.e. make a call, send an SMS, use data, etc.) for a period of three months but does receive calls on the line, has he or she churned? Or, if a postpaid subscriber pays the subscription fee for six months in a row but has no inbound or outbound activity, has he or she churned?

Driving variations in the ways telecoms define churn are several factors, such as telecommunications authorities with their own definitions stepping in, subscribers who may be more mobile / go abroad for extended periods of time in a given market, competitive landscape, etc. To effectively define churn, telecoms need to conduct analysis regarding their consumer’s behavior, identifying at what point of behaving in a certain manner a subscriber does or does not leave.

Analyzing past churners to identify the number of weeks / months of inactivity preceded their leaving will help nail down a definition around churn, allowing for a standardized view across the organization on this issue. This definition will in all likelihood be different for prepaid and postpaid customers, and will serve as the driving measurement principle going forward.

2. Build Predictive Churn Model

A predictive churn model, for those who may be unfamiliar with it, is a tool that helps telecoms (or companies in other sectors as well, for that matter) identify which of its subscribers will churn within a given coming time period. If a telecom loses 100,000 customers each quarter, for example, the predictive churn model tries to identify who those 100,000 subscribers will be in the coming quarter.

Without a predictive churn model in place, proactive retention efforts cannot be put into action. It is this tool that defines which customers to contact proactively, in order to prevent the churn occurrence from happening. Otherwise, a mobile operator would have to guess which subscribers may churn, causing severe inefficiencies in retention efforts as well as yielding little results.

A churn model must be tested before it can be relied on to support retention efforts, as it may not be predicting accurately enough which subscribers may actually churn. If the model, for example, predicts that 10,000 of a group of 20,000 subscribers will churn, but two months down the road only 5,000 have churned, the model has a 50% error rate, essentially unusable. As an entire proactive retention effort will be designed around the models, it is critical that their accuracy be tested and ensured before going live.

3. Design Offers & Channel Strategy

Based on the findings coming out of the predictive churn model, the next step will be to design offers to be made to the various groups predicted to churn. The variables to be considered in designing the offers are as follows:

  • Churn Likelihood
  • Customer Product (Prepaid or Postpaid)
  • Customer Value
  • Customer Usage Behavior (Local / International Minutes & SMS, Data, VAS, etc.)

Utilizing these variables, a dozen or more subscribers groups for targeting should be defined (i.e. postpaid subs with ARPU 500+ USD and above, with 25 – 50% churn risk, with 50+% of MoU international). Based on the defined groups, the next step is to design customized offers that appeal to their unique characteristics.

It is critical that the offers be relevant to the groups; using one or two standard offers won’t cut it, as an unattractive benefit won’t do much towards keeping the customer from churning. For example, an offer to a prepaid subscriber whose ARPU is 30 USD, has a 25% likelihood of churning and only sends SMS in terms of behavior must be drastically different than that offered to a postpaid subscribers whose ARPU is 150 USD, has a 50% likelihood of churning and is active across the board, using data, VAS, etc.

The offers to be used should be of the kind that at least ensure the subscriber stay for a certain amount of time, preferably of the kind that even drive up ARPU. In its simplest form, the offer could be one that gives the customer a certain amount of free minutes each month for a certain number of months. In a more complex form, it could be one that gives a discount on a fixed postpaid plan fee for a certain number of months if and only if the subscriber migrates from their prepaid plan (this would be an example of driving up ARPU). One-off offers should be avoided, as the subscriber can churn immediately after depleting the benefits (i.e. 50 USD discount on a bill this month). One additional note – loyalty program benefits can also be used here, giving the subscriber additional points or services through the program for a certain period (i.e. if you stay with us, we’ll give you access to our platinum level service of our loyalty program for the next year).

Once a primary offer is defined for each target group, a secondary offer should also be defined, in case the subscriber rejects the primary offer. Not all offers will be attractive to all subscribers in a given group; having a secondary offer will ensure higher retention rates are realized through churn prevention efforts.

The above, of course, must all be in line with telecommunications regulations in the marketplace – in some markets, such offers must be made available to everyone – in others, such below-the-line offers are allowed.

Once the offers are designed and ready in the billing systems, the next step is to determine the contact strategy – who will be contact by what channel in what order. Using a contact center is the most effective method in retaining subscribers, but doesn’t make financial sense in many cases – a breakeven analysis will help determine which subscribers should be called (i.e. only those subscribers who have over 100 USD ARPU and have a 25+% likelihood of churning). Those not to be called can be contacted via SMS, e-mail, or auto-dialer (again, regulations need to be reviewed here).

4. Prepare Operations & Pilot Efforts

The final step, before going live, will be to design the enablers to drive the retention efforts – including alternate scripts, SMS text, training materials, etc. The importance of these cannot be stressed enough – a poorly trained call center representative with an ineffective script in his or her hands may cause more harm than good. The piloting phase is a perfect opportunity for examining the effectiveness of the materials as well as the agents.

At this point a critical fact about proactive retention efforts must be mentioned – such efforts only work when the retention offers are made mainly only to those subscribers who are about to churn. Even the most sophisticated and advanced predictive churn models in the world are unable to only pinpoint churners; that is, a model is able to state that 50 of a specific 100 subscribers will churn, not pinpoint the specific 50. The trick lies in making an offer to the 50 who will churn, and not to the 50 who won’t. If offers had to be given to all 100 subscribers, the business case around proactive retention efforts would not make sense.

As such, call center representatives need to have a target around the percentage of subscribers they can make offers to – thus, their goal is to gauge the satisfaction level of the subscribers during the conversation to determine whether or not to make an offer. This can be done with excellent scripts and effective training, arming the representative as much as possible to ensure the desired outcomes are realized.

The final step, before going live, will be to pilot the churn prevention efforts, to ensure the optimal strategy is in place, taking into consideration several different factors:

  • Reach Rates, by Channel
  • Response Rates, by Channel
  • % Satisfied
  • % Accepting Primary Offer
  • %Accepting Secondary Offer
  • % Rejecting All Offers
  • % Saved
  • % Change in ARPU

The above findings must then be used to fine-tune the offers, the timing of the offers, the scripts, the training materials, etc. More importantly, the impact on the bottom line can be assessed, to determine what channel strategy makes sense, helping fine tune which customer is contacted by what channel (based on response and retention rates as well as % change in ARPU).

Mobile operators that haven’t yet launched their own proactive retention efforts should seriously consider doing so at this point. Such strategies are proliferating through the telecom industry, as the challenge of retaining subscribers is increasingly becoming more and more difficult.

Tags : analytics, call center, consulting, data mining, fact-based, intelligence, marketing, program, retention, strategy