Customer Segmentation for contextual marketing in Telcos

CSPs have been using customer segmentation for many years now – demographic, usage based and ARPU driven.  In the times of voice and SMS, such segmentation sufficed the purpose of selling these basic services.  With data explosion and revenues increasingly influenced by data and VAS, such segmentation is not fit for purpose and needs to change.

In addition, CSPs have granular customer interaction data passing through their network – Apps used, URLs accessed, locations and times for access, type of network of access, recency and frequency of access and so on.  A lot of these contextual data is extremely valuable for segmenting customers in a different way – using their interests derived from these data.  Some CSPs have started exploiting these data for segmentation but there are challenges ahead in terms of technology, data privacy, derived interests and more importantly potential business benefit in terms of uplift over existing campaigns.  I would like to dwell on some of these aspects in this article.

1.  Technology Issues

Much of the traffic today on the mobile internet is transitioning to HTTPS.  With some estimates, more than 60% of traffic is HTTPS.  Customer conscious of their privacy, are increasingly using VPN services to access internet.  Much is the URL and APP data gathered by CSPs become not so useful when these connections are encrypted as CSPs lose the long URLs they used to get before encryption.  These long URLs allow them to derive customer’s intention and hence interest.

Second challenge is to classify multitudes of URLs and APPs into meaningful categories and interests that can then be used for further segmentation.  There are organisations such as similarweb and ZveloDB that offer classifications.  However, output of these classification engines is not of great quality because input are the high level URLs due to encryption and HTTPS issue discussed earlier.

So quality of category data presents the first challenge to segmentation.

2.  Cluster Dimensioning Issues

Each category/subcategory output from categorisation tool becomes a potential dimension and these can go up to as high as 300.  With recency and frequency values over 300 dimensions, the input becomes quite complex and cluster output quality can be poor.  So concepts from text analytics might have to be used to manage the dimensions and also preprocess data.  Concepts of TFIDF – Term Frequency Inverse Document Frequency can be applied to dimension matrix to remove rare or more frequent terms and then balance the remaining terms as per their frequency across all corpus.  Such processing is shown to to give much high quality cluster output.


3.  Creating Personas or segments from clusters to make sense of them

When you create clusters, sometimes it is difficult to describe these clusters to the business community.  Supervised classification using clusters can create rules that can help define clusters.  Another challenge is to integrate existing demographic, psychographic and value information with clusters to create personas – way to describe segment fully in terms of their characterisitics.  Since demographic and other information is available on sample basis, it is tricky to merge it with cluster information to create personas.  Additional market research for the customers in the sample for clustering might be needed to integrate this information.

4.  Data Privacy Issues

CSPs and OTT players are governed by different laws with respect to privacy.  For example in the US, FTC laws govern Facebook and Google while FCC laws governs CSPs.  FCC laws tend to be tougher on privacy than FTC and hence OTT players have some advantage of being able to use your data vs CSPs being able to use data for upsell and cross-sell or even making content free through ad-supported sites.  I think customer education with respect to privacy and what data CSPs can really see versus what Facebook and Google know about you might help CSPs to convince its customers to share consent.


5.  Return of Investment

We all talk about personalized offers generating huge benefits – Netflix, Amazon have proved it many times. But there is little evidence of CSPs achieving substantial lift through such behavioral segmentation over their existing baseline segmentation based on demographic and service usage.  The key challenge is the attribution where it is tricky to measure and attribute incremental value to such behavioral segmentation and not to other aspects of the business that are changing at the same time.  Sophisticated econometric modeling is needed to measure and quantify attribution.


Despite these challenges, there are huge opportunities for CSPs to make use of their data to create meaningful segments based on customers interests.  In addition, they have huge advantage in terms of customer context – device, location, time, network presence which can be critical for contextual marketing in addition to clustering.


I feel that in the coming months, we can see lot of action from CSPs on this front – Verizon with Yahoo, Telenor with Tapad, Telefonica with Amobee and so on.


Analytics in the Telco Context

Big data analytics has been hot topic for some time now and like everyone else, I thought I needed to understand it better from telco perspective and sieve through the hype to understand the reality myself.

So I proceeded to do analytics certification at Wharton Business School recently and their classification of analytics was pretty good.








If I see current Telco focus, it is in the top left half of this table – Customer and Operational Analytics in terms of their descriptive and predictive capabilities.  So I will focus on these two in my current post.

Telcos are doing a lot of descriptive analytics on the customer side – customer segmentation – demographic, psycho-graphic are main focus.  However, with the use of granular service level data per customer, some operators have started to look into lifestyle or interest based segmentation with RFM (Recency, Frequency, Monetory Value)  based clustering.  These lifestyle segments are richer than existing segmentation and will be used to enhance current customer profiling.  Technology is also enabling the use of cross-device behavior analysis – companies like Tapad are leaders in this field.

Predictive customer analytics is where Telcos are still scratching the surface.  At present, most are occupied with churn prediction but very few are dabbling in the area of machine learning and CART techniques to use lifestyle segments to predict propensity to accept targeted content offers.  Prescriptive analytics connects predictive analytics to maximisation of business benefits.   There is lot of scope for automation and machines learning in prescriptive analytics and telcos are way at the bottom of the learning curve.

The area of operational analytics is much less explored by telcos.  Today, very basic descriptive analytics on operations is carried out, mainly on the network operations side.  Some analytics is carried out on the customer care operations as well.  However, the real benefit in network operations is predictive analytics that can anticipate network and service problems before they occur so prescriptive analytics can applied.  For example, analyzing alarm data and performance management data  and using regression techniques to predict service quality degradation for future traffic growth can take proactive actions to fix before it happens, thus saving huge amount of firefighting efforts in the future and protect network quality brand.


In future, actions can be improved through machines learning, analysing impact of past actions on service quality and automatically recommending changes to future actions.  this area of prescriptive analytics will generate the biggest benefit for telecom operators.

However, as many believe or suggested, predictive and prescriptive analytics alone are not going to solve all problems of telcos.  Associated processes and Governance has to change to act upon the insights from analytics.


I am very excited about the use of predictive and prescriptive operational analytics and look forward learn many more interesting use cases.

Business to Business CEM – very different that B2C but very important

When we talk about CEM, we mostly give examples on B2C space.  But lately, there has been lot of interest on B2B segment and some CEM conferences have been held specifically focused on B2B CEM.


I have been struggling to understand B2B CEM space and recently got an opportunity to perform CEM maturity assessment for a converged operator in Europe, struggling with customer experience with their large customers, especially when they wanted to new converged offering to their customers, upgrading legacy solutions.  Here are my key takeaways from the the conference –


  1.  Business Continuity, Productivity and Security are top requirements in B2B CEM
  2. Understanding buyer personas – which people are involved in buying process, what are theirs goals and preferences, how end user is not involved in the buying process
  3. Most B2B customers want some SLAs for their CEM but are not willing to pay for it.  But they might consider paying for additional services such as Security, Analytics
  4. Persoalised service and channels of engagement are important to consider when designing B2B CEM

Orange Business Services shared their research with IPSOS on B2B CEM –

  • Business Continuity in mind
  • Ease of doing business – New Service Activation, Support, Self Care (Turk Telekom mentioned that it took 19 steps to change speed on a package!)
  • Value added services such as Analytics
  • Security of critical business resources and processes
  • Increasing Productivity of business

Vodafone raised issues of procurement for SME and large corporate customers.  Most interactions are with plan administrator and commercial with end users not getting represented properly.  The disconnect between admin and end user create problems when finally service is launch as end users are not aware of service catalogue, SLA levels and support agreements.  So understanding buyer personas and designing CEM fo them is pretty important.  A book Buyer Personas describes in details about buying process.

In terms of channels of engagement, large corporates have a dedicated account manager but he/she doesn’t have information on the customer experience or the SLA of the customer.  If you can’t measure it, you can’t manage it.  For SMEs, collaboration with local partners known to SMEs in a given area should be considered.  Local partners knows more about localisation of service.  Shop within a shop for SMEs might be considered as done by Proximus in the Netherlands.

Very interesting discussions on the SLAs – most customers won’t pay extra for SLAs as they have higher expectations from supplier.  In fact, a few customers mentioned that they would put CEM SLAs as a condition of contract.  So SLA offering and management for CEM is not longer a value added service….

For vendors supplying CEM, internal organisational changes are needed to better serve B2B customers.  TDC Denmark has developed industry specific teams to serve its B2B customers.  SoHo and SMEs are integrated into Proximus’s consumer business to better serve them.  Internal awareness of B2B CEM SLAS through Digital dashboard can be improved as tried by Turk Telekom.

In essence, B2B CEM is different than B2C CEM and operators have to make changes in their approach, organisation and channel engagements to better serve B2B customers.




How bad multi-channel experience ruins the brand?

All brands, in all verticals, are getting into multi-channel customer engagement due to customer’s multi-touch journeys across channels and not doing so will give competition advantage over you.  However, doing it wrong will cause more damage to you than not doing it at all!


My recent experience with American Express in India is a case to this point.  I have been Amex platinum customer for past 6 years and had some bad experiences with them – bill sent to wrong address, bill signed by security person not authorised to sign, food offers not sent on time and so on.  But they are on their way down in terms of customer experience and dealing with their customer complaints, without empowering employees to do anything about it.  This time, I called their contact center asking them about this year’s food offers.  First thing, I had to go through multiple security hoops despite validated on IVR.  Second, Amex told me they discontinued coupons to go green.  Though this is noble cause, not communicating to customers about this change of policy has done more harm to them.  Third, customer service asks me to go online again to search before I explicitly asked them to help me.


The website is so badly designed that it took me 5 clicks to see offers page.  Search function doesn’t work and typing “bangalore” didn’t result in any output.   Response to email complaint was done post 48 hours in which my name was spelled wrong.  So what does this tell you?  American express has not thought about proper multi-channel customer journey.  they haven’t designed online channel well and hence migrating customers to this channel for offers resulted in poor experience.  Since I went to online then to contact center and again contact center called to apologise, cost to serve went up, defeating the whole purpose of online.  Most important, I become their detractor now from promoter….that is bad given I spend at least $4000 a month on their card.  When I asked them to deliver a new card, they promised 2 days but card didn’t get delivered as their courier company couldn’t locate my address – this despite the fact that same courier company delivers my statements on time every month!


So how to fix this mess?  One, design customer journey and understand key moments of truth, across channels.  Process design alone is not enough as one can’t anticipate all permutations and combinations of customer demands.  So empowering front line employees is key to make decisions on the spot.  Third, clear, concise and timely communications with customers is critical.


I hope AMEX listens and changes!


Can a bad experience at moment of truth make you a detractor?

Recently on my trip to Bangkok, I had one of the worst experience in a 5-star boutique hotel.  I didn’t have 4 bottles of water I always ask, BBC channel was not working like last time, and I was made to wait at breakfast for 20 mins before I was even served a cup of tea!  In addition, AC stopped working and bathroom smelled bad.   It made me wonder whether I would go there again despite several goodies I received after I complained about it…


It made me think.  In my stay with hotels, what were the things for which my expectations were high and not met, and which are the things where my expectations were low and not met?  Also, is this repeat experience in near past or distant past?


I think when expectation are high, these factors are hygiene factors.  Good experience on hygiene factors is not going to make me wow and unlikely for me to tell 10 others.  These factors vary for different personas.  Certain factors are common for all personas – clean room, clean bathroom, working AC or heater, decent breakfast.  For business traveler and premium member like me, fast check in and check out and access to business center are hygiene factors.  Personal recognition ( recognizing by name as I am frequent visitor) and specific preferences in the room such as additional water and fruits are also hygiene factors for me.  Things such as complementary booking to a restaurant or a show, complimentary limo pick up and drop off are delight factors where my expectations is low.


Now if you don’t deliver on moments of truth for hygiene factors, customer will get annoyed.  But if he is associated with you for decent amount time and he/she likes your brand, he or she would give constructive feedback and expect some concrete actions from the brand to fix.  He or she would give brand one chance.  But if he or she experiences similar issues on these hygiene factors repetitively and in short duration of time span, memory hasn’t faded and negative impact has compounded now, causing him or her to become detractor.


So repetitive bad experience on hygiene factors over a short memory time frame is the worst and brands must avoid it.  But how to ensure this?  One can do journey design, identify moments of truth, create metrics to measure, define actions and processes and so on.  But we can’t anticipate every situation, every permutations and combinations.  So empowering front line employees to take actions on their own good judgement is key.  This judgement is guided and framed by company’s Customer Experience strategy and culture.


I plan to give my hotel one last chance….else I am gone forever.

CEM Value and Outcome based business models – theory vs reality!

In his seminal book “The strategy and Tactics of Pricing”, Nagle and Hogan stressed the importance of Value based pricing and economic value estimation.  It emphasizes on the fact of value creation, value communication and value pricing to increase customer’s willing to pay and finally they actually paying for it.

All solution selling is difficult compared to box selling and CEM solution is even trickier as value is not clear.  What I mean is that value attribution is not clear.  Cost efficiency improvement or revenue increase can’t be attributed fully to customer experience management initiative.  There are so many other dependencies that very few use cases can show incremental benefit.

In order to resist discounting, CEM vendor have been developing value argumentation and business outcome calculators to convince the operators about tangible benefits of CEM solution.  CEM value modeling shows different operational metrics being influenced by CEM such as First call Resolution, Mean Time to Resolve (MTTR), reduction in churn rate, incremental data volume and its potential monetary value, and so on.  Operational metrics are then mapped to business metrics such as revenue and CLV and CE metrics such as NPS.


This is all great and right way to go about it.  However, when it come to discussing business model where payment is based on outcome, i.e. risk and reward model, same vendors who were pushing value based pricing back out.  I have seen in many deals over the past 3 years in the CEM space.  There are multiple reasons for this.  First, value modeling is paper exercise with many measurements not validated with rigour.  For example, it is not possible to measure improvement in first call resolution on specific calls due to practical challenges of measuring it, incorrect tags of calls and so on.  Secondly, more the measurements, the bigger the challenges of Governance.  Disputes and arguments regarding measurements and attribution are bound to happen and it creates additional overheads both operator and vendor.  Third, lack of strong political support for front line sales people from their powerful bosses at HQ.  There is no incentive tied to profitability in addition to revenue targets.   Procurement wants mechanism that is easy to audit and enforce.  Last but not the least is the control.  Vendor committing to outcome based payment wants full control of network and IT operations to implement actions from CEM insight.  One vendor having end-end network and IT operations is seldom the case.


So what is the way forward?  In my experience, limit the measurement criteria to one or two business metrics that are at business level and may be one CE metric.  For example, for operator launching 4G to improve data services, number of unique data subscribers, Data ARPU and NPS can be metrics to consider.  Second, operators just want to share risk and not reward.  No vendor is going to accept this deal.  So for win-win situation, both risk sharing based on outcome, but reward based on exceeding outcome also has to be part of the deal.  Governance is key.  For both vendor and operator, trying this mechanism on a limited scale first can uncover the realities of working in this new model and make adjustments.


However, I do feel that outcome based business model is the way to go for CEM deals!



Net Promoter Score – Driving factors and their cross-correlation!

A lot has been talked about Net Promoter Score – its benefits, its simplicity and its relationship with revenue growth.  So I won’t discuss these topics again but something different that has been bothering me for some time.

I worked on a consulting engagement for a telecom operator to improve their NPS – in particular, my task was to understand underlying factors for their detractors.  So I delved into analysing customer feedback data from the customer feedback management system and found 3 primary reasons for detraction – Network Quality, Customer Care, and price.  60% of the detractors cited network quality as the main reason for their unhappiness.  So it was obvious to focus on Network Quality and understand actual factors that were responsible for poor network quality.

Follow up questions with customers shed more light and we derived 3 factors for network quality – Mobile broadband speed, Voice quality and indoor coverage.  Based on best practices learned in NPS certification, we collected objective data on the key quality indicators for these factors and performed correlation to determine what factors are strongly correlation to NPS and hence first to focus on.  It turned out Mobile broadband speed and indoor coverage had strong correlation and voice quality had weak correlation.  For customer care, not able to resolve the complaints first time and time to resolve the complaints were main factors – they showed strong correlation with NPS.

We proceeded to improve network quality through targeted optimisation of speed and coverage and hoped to see reduction in detractors in 2 quarters, NPS being the lagging indicator.  What we saw was surprising – we expected Network detractors to go down and hence overall NPS to go up.   We did see that 50% of people were detractors due to network quality compared to before – 20% drop, but overall NPS didn’t improve by much  – a mere 1 point.    So when we performed detailed analysis, we realised that now detractors due to customer care have increased and this had offset the reduction of detractors due to network and hence overall NPS didn’t change much.  In fact, we expected more sharper drop in no of detractors due to network.

It turns out that different factors of NPS – Network quality, Customer service and price have strong cross-correlation between each other.  Customer Care agents were not able to handle network related queries and that generated dissatisfaction not only on customer care but also on network quality!  So gains from one factor can be compensated by bad performance in other aspect.  This finding was also confirmed by a research paper from Indian Institute of Management, Bangalore in a study of customer satisfaction for Mobile Network Operators in India ( Source: Structural Equation Modeling of determinants of Customer Satisfaction of Mobile Network Providers, October 2014)

So it is important to set expectations.  CTO for the said operator was under great pressure to improve network quality so that overall NPS score will go up.  But even after improving Network quality, it didn’t result in overall NPS improvement as well as network NPS improvement as per expectation.  The lack of investment in Customer Care while focusing solely on network quality didn’t yield desired results.

So understanding cross-correlation between different factors of NPS is important to assess impact on actions on one factor and neglecting the action on others.


Banding together Network & IT for CE Journey experience

For Telcos, the bread & butter is network and this is where all the CEM initiatives start and rightly so.  One has to have best network irrespective of the technology used.  If there are network issues, no point talking service quality and then customer experience indicators.


However, once you fix network part, the game changes drastically.  When you start thinking about customer journey, you realize that many interactions of customers with your brand involves IT systems. In fact, I will go on to say that IT dominates interactions with customers except when he or she uses Telco’s traditional voice and data services.  Even with Apps, many of the interactions of customers are happening with IT systems – Telco’s internal IT systems and increasingly with those outside Telco domain – with 3rd party players in the ecosystem.

In order to assure consistent superior experience, CSP first needs to measure both network and IT experience at as service level and at a customer level across a customer journey.  Without such a visibility, CSP won’t know the real end-end experience of the customer in a business transaction.   In order to measure Customer’s journey experience, CSP has to be measure experience at different parts of Customer lifecycle and then join it in an intelligent way to get end-end visibility.  TM Forum has publicized Frameworx document that has CEM lifecycle framework – Buying, Using and Sharing.  From a mobile customer’s perspective, buying means – research on multiple channels – mainly online and then buy at the store.  Using means calling and using apps.  Getting helps means calling contact centre or increasingly look online for help or use self care apps.  Sharing means writing on CSP’s facebook page or to your friends.   If the goal is to get positive sentiment and eventually make customer a promoter, then experience at all the stages of the life cycle should be good and consistent, with some experience surprisingly good.


In order the assess this experience across life cycle stages, getting data from multitude of IT and network systems is first step.  To get data from IT systems, APM vendors such as Splunk, Compuware, Riverbed can be used.  They use soft probes and/or log correlation mechanisms to create a business transaction at the customer level.  It is important to assess these vendors from the perspective of per user transactions and NOT Application performance.  This IT data can be attributed to buying and getting help phases.  For use phase, data from probes is required to get similar per service per user visibility.







The experience in each of the life cycle stages can be quantified through sophisticated data modeling.  The challenge is how to correlate data from IT systems related to buying phase to network data from the use phase.  Detailed understanding of data model in IT systems and network DB would reveal potential connections that need joining to create a single view of customer.   Once the links are identified, IT and network transactions are stitched together to create business transaction that is stored in data base with required granularity.  Each stage of life cycle should have its quality of experience score based on weights of key measurements.  For example, for buying journey, waiting time in store, interaction time, successful activation of required services, time for activation of services should be considered.  For use stage, network experience measurements such as speed consistency, first page rendering time can be considered.  For help stage, no of calls to help, first call resolution, mean time to resolve should be considered.


For overall experience, additional weights for each of theses QoE scores should be considered.  These weights can be gathered by integrating subjective feedback from customers – customer satisfaction survey or Net Promoter score.  Mapping subjective feedback to a set of measurements and using regression techniques, factors or weights for each of the KQIs and quality of experience scores can be derived.  Once these factors are know, overall customer experience index for entire customer journey can be derived and tracked for all of CSP’s customers.


Once the Customer experience Indicator or equivalent measurement is tracked, then it can be used to assess overall customer journey experience and can be used for demarcating problem if it shows degradation.  In one such implementation, I have found that CSP didn’t have any visibility of their customer’s experience – in their store, on their network and also on the contact center.  After implementing such solution, CSP was able to assess experience of their newly joined customers.  For example, CSP found out that 9% of their customers left the store after 10 mins of waiting time.  CSP had this visibility on each of their store location and time of the day,  Armed with such visibility, CSP can take corrective actions to fix the issues and improve customer experience and perceptions.






Telco Network NPS – Should you do it, what to measure and how to action it?

Net Promoter Score or NPS is being increasing used in many organisations including major Telcos as a transformation tool to change their organisations to respond more faster to customers and make customers loyal.  NPS economics is well known – Loyal customers last longer with you, cost to serve goes down, they buy more and also refer more customers.  So it impacts not only the revenue growth but also the bottom line.

For telcos, Network is core to the offering.  From customer point of view, it is more than a touch point.  Customers use network all the time to do what they want to do.  So treating network as a touch point is doing injustice.  Of course, there are other touch points that are really touch points – retail store, web, customer care centre, and other self care channels.  So it is more appropriate to use Relational NPS to understand the customer view on the network.   But, relational NPS wont’ give you clear information about the drivers for loyalty – what exactly in the network that promoters like and distractors dislike.  It is also important to know the importance of the drivers.  One can perform gap analysis between importance and satisfaction to focus efforts to improve .  But it is also critical to understand the correlation of drivers to NPS.  Drivers where there is a gap between importance and satisfaction but low correlation to NPS is lower priority than one where correlation is high.


One middle east operator carried out network NPS survey and found very interesting insights – most customers had low importance on mobile broadband speed but high correlation to NPS.  This is hidden driver – but for this operator, gap between importance and satisfaction was low and hence operator decided to divert efforts on improving speed to some other driver.  In their case, voice quality had high correlation but big gap between importance and satisfaction.    So CAPEX and OPEX was spent on improving voice quality and NPS after 2 quarters showed significant positive movement.


NPS is finally a sample and hence operators have asked me interesting question – how can I use NPS to understand the impact of driver on all my customers.   This is where regression methodology comes into play.  Let’s assume that Mobile broadband speed is an important driver from NPS survey – high correlation to NPS.  But, gap between importance and satisfaction is huge.  So operator decided to focus on improving mobile broadband speed.  But, how does operator know what exactly to improve and where to improve first.


This is where customer comment needs mapping to a measurement in the network that we can measure reliably for every customer and every service.  For example, speed can be first page loading delay, standard deviation of average speed for a popular website ( consistency), subsequent page load time, etc.  So here is the process to follow –


1.  For every customer who has filled NPS survey, find network usage data for at least past 3 months if not more.

2.  Get data on speed on every session, page load delay and other parameters that you think are relevant from customer point of view when he or she talks about speed.

3. Create a relationship with NPS as output variable and speed related parameters as dependent variables and carry out regression equation.


4.  From regression equation, you will know which variable is significant and which is not.  You will also find out the potential impact on NPS of improving the variable ( regression coefficient)

5.  Now you can plot NPS for all your customers and figure out clusters where improvements in speed is needed that can have the biggest impact on your NPS.

6.  Now focus your investments in these areas and on these variable to improve.



To understand the drivers, one needs to have some additional questions to the typical 2 question NPS survey to get detail subjective feedback and also to get data for driver analysis.  Of course, response rates should be taken into account as survey gets longer.


So I think network NPS is a good idea, provided it is not done in a pure transactional way and driver analysis is carried out to close the loop for issues that will have maximum positive impact on NPS.

Net Promoter Score for CEM

Net Promoter Score or NPS is gaining momentum especially in the telecom sector with many operators using NPS.

However, most operators are using only the relational part and then mixing the transactional NPS into it.  Also, many operators are using NPS in isolation and not with customer journey concept.  So NPS is not actionable.  Having discussed and seen NPS implementation in many Opcos, here are my thoughts –


1.  Opcos have to use transactional NPS to measure key customer journeys that they intend to improve.  Each journey have its own additional CE metrics (outside in) and functional metrics ( inside out).  Sometimes, it is impossible to define and measure CE metrics per customer.  Here, NPS comes to help.  However, NPS should have follow up questions – not more than 3-4 to understand the reasons behind promoter ratings as well detractor ratings.

2.  Journey framework provide a structure to assess customer experience.  Hence, NPS usage at only a few touch points is not useful.

3.  There is need to identify drivers for transactional NPS and their impact on NPS.  By doing so allows you to forecast potential NPS improvement as a result of certain actions and also quantify the improvement in terms of Customer Lifetime Value.

4.  Multiple Transactional NPS are mapped to relational NPS that aligns with a brand.  Hence, relational NPS is important to judge overall organisation pulse during the CEM transformation.  But relational NPS is not actionable.

5.  It is important to reduce detractors first than increasing promoters.  Negative word of mouth is more damaging than positive words of mouth.  Bain, and London School of Economics research both point to much stronger correlation between negative word of mouth and decreasing revenue than positive word of mouth and increasing revenues.

6.  It is critical to map relationship between transactional NPS and operational metrics to understand the impact of organisation improvement on NPS.

7.  CE metrics should not be abandoned when NPS is launched.  NPS is finally a sample while CE metrics are measured per customer for all your customer base.  There should be a mapping between CE metrics and NPS to confirm that NPS score does indeed reflect customer experience across all the customer base.