Analytics and Data Mining,ROI,Social CRM

Data-Driven Social CRM

15 Apr , 2010  

I have been reading some very interesting books about strategies around becoming a customer-driven organisation and also about the benefits of Customer Engagement Programs (more on that in later posts). In the discussion around Social CRM we agree that it is beneficial to business to engage the Social Customer, but what seems to be more difficult to articulate is the Business Value Proposition – why should a company invest in Customer Engagement Programs in conjunction with CRM? What value is in it for the company and what is in it for the customers? We also talk about the need to move from value-in-exchange thinking to value-in-use. What tangible benefits are there for companies to take a a customer Lifetime Value approach rather than concentrate on the sale? Or simply, how can Social CRM Strategies can provide a significant impact to the bottom line.

This is where Jim Novo’s book “Drilling Down” got me thinking – hard (and kudo’s to fellow AC-er Mike Boysen for helping me shape my thoughts). Jim Novo’s premise is that the data you have in your systems about your customers will allow you to segment them based on their behaviour and not sociodemographic characteristics and concentrate your resources on those that will bring you the highest value. It sounds kind of like inside-out thinking, but believe me, it is quite the opposite. Rather than concentrating on attrition and win-back of all your customers(which is far more expensive than customer retention), the method he proposes also allows you to pick up on deviation from expected behaviour, use these as triggers and perform what I’d describe as “just-in-time” retention activities to keep the patronage of high-potential or high-value customers.

This excerpt is taken from the free chapters of Jim’s book (you can get them here by signing up for his newsletter):

1. Past and Current customer behavior are the best predictors of Future customer behavior.

You can predict future behavior based on an understanding of past behavior, and use this knowledge to improve marketing or service programs [based on] actual behavior, not implied behavior.

2. Customers want to win at the customer game.

They like to feel they are in control and smart about choices they make, and they like to feel good about their behavior. Marketers and service providers take advantage of this attitude by offering programs and communications of various kinds to get customers to engage in a certain behavior and feel good about doing it.

3. Data-Driven programs are about allocating resources.

Data-Driven marketing and service programs are among the very few allowing you to accurately measure ROI. [It’s about] reallocating capital with low return to higher return projects or programs, generating higher profits in the process.

4.Action – Reaction – Feedback – Repeat.

Data-Driven marketing and service programs are driven by creating continuous communications and interactions between the business and the customer, and analyzing these interactions for challenges or opportunities.

This is the example Jim uses to illustrate his propos:

For example, a win-back program is triggered when the customer defects. Have you switched long distance or cellular providers lately? Did you get inundated with win-back calls begging you to reconsider? “Jim, we just wanted you to know we have lowered our rates.” Yeah, well, thanks for telling me after overcharging me for the past six months! But could they have known I was about to switch by looking at my behavior?

Sure. If they had looked at the calling patterns of previously defected customers like me, they would have seen a common thread in the behavior.[…] The proper profit maximizing approach is to wait until I look like I’m going to defect, and then call me and offer a lower rate before I defect.

Too late. The customer no longer is one.

The “Drilling Down” approach is aimed at constantly improving and optimising the value of both the company whilst ensuring a satisfactory customer experience.

What I like about Jim’s approach is that they provide a concrete Business Value Proposition for implementing a Customer Retention Strategy that optimises the allocation of the company’s resources by focusing on the most profitable segments – or in Finance speak: optimize the Return On Investment.   And as you probably know, the quickest way to a CFO’s heart is not though his stomach, but through the figures…The only issue is that it requires the company to move away from a quarterly quota model to a customer Lifetime value approach (see the book for more details on this last remark…)

Enter The Social Customer

In Jim’s example above he focuses on someone who leaves and moves on without looking back. What has now changed the game is that the Social Customer – through her interactions with her peers – can have an impact on the cost of new customer acquisition and as well as on retention, both in a positive and negative sense. No longer is the customer ‘an island’ because of ubiquitous opportunities for peer to peer conversation. These can influence customer behaviour patterns on a world-wide scale and thus adds a variable to the equation to the above that needs to be taken into consideration.

In a Social CRM strategy, to address the Social Customer whilst at the same time optimising the company’s resource usage and profitability, key to me will be the ability to link the conversations of your customers to the segments you have identified of high/low potential/value. This means that rather than tracking all conversations everywhere and pumping these into your database, you should be focusing on your high potential/value customers Furthermore your understanding of the customers should be aimed at identifying behaviour patterns and triggers within the channels where your customers express themselves and that you can monitor.  The barrier to pick up the phone [or fill in whatever channel you like] can be high for some customers and this can deprive you of the trigger needed to track a deviation and react accordingly. The conversation the Social Customers are having through these public communication means may now provide you with this trigger that you would have missed out on before. And rather than allocate your resources to responding to every person that wants to rant, you can concentrate your efforts on those that you would like to retain (whilst remaining ‘polite’ with those that are not of direct interest of course).

Although I realize that this line of though does not address all the opportunities of Social CRM and Social Business Strategies, it is my first attempt at rationalising  and articulating the Business Value Proposition. With a broad stroke of the pen I have put other elements such as peer to peer support, ideation, customer experience programs, open innovation etc. under the header of Customer Retention Strategies, and I do apologize for this.  I will try to work out my ideas in further posts.

What do you think? How would you formulate the Business Value Proposition for Social CRM and Social Business.


26 Responses

  1. Wim Rampen says:

    Hi Mark,

    It makes perfect sense to focus your efforts on your most valuable Customers and not all customers. It remains difficult though to assess who your most valuable customers are. Profitability is still the most used metric for this, or Lifetime value for the advanced ones. It remains to be seen if your most profitable customer is also your most valuable Customer. Last year we touched briefly upon the subject of network value and referral value. These elements are not well understood, imho (but I did not do very deep research into these metrics), and certainly not widely applied.

    With more clarity now on what Social CRM is about and what is at the core of it (empowering your Customers to create value for/with the Customer, herself or others) this is a good time to push that discussion to the next level. Because only if one can truly understand who the most valuable Customer is, one can understand “how come”, which is a key-(actionable)-insight one needs to focus a company’s efforts.

    What do you think? Any thoughts/ideas/examples of Customer referral and network value? Are these the right metrics?

  2. […] Data-Driven Social CRM Published: April 15, 2010 Source: Social CRM ideas by Mark Tamis I have been reading some very interesting books about strategies around becoming a customer-driven organisation and also about the benefits of Customer Engagement Programs (more on that in later posts). In the… […]

  3. jayhawkscot says:

    Hi Mark,

    In its purist sense, alot of what you say I right, and jives well with Esteban Kolsky reality check post on TheSocialCustomer,, especially in terms of the quickest way to a CFO’s heart.

    The Drilling down approach is tradtional CRM in its purest sense, although I have some questions/doubts about its ability to “ensure a satisfactory customer experience” using behavioral data.

    Concerning the transition to the Social Customer space – that’s where the complexity comes in. Tracking my high-value customer conversations in social space assumes that 1) I can link these, 2) I can link these in real time, 3) I can analyze and react in real time (or at least before it’s too late, and 4) that the social conversations provide enough granular data about a potential change in behavior (ie; using Jim’s example of long-distance cellular providers – a high LTV customer asks a community of trusted peers about a competitors offer he/she just saw and what do they think of this competitor and their product / service, I see this in real time and provide a better offer). Great in theory, but, imho, there are currently so many holes in the reality of this…

    You provided a great way to clarify the concept of Social CRM with tradtional CRM in Finance speak, and as Esteban warned us, this is reality.

  4. Mark Tamis says:

    I agree with your pov about not knowing whether your most profitable customer is also your most valuable one. “Drilling Down” tries to provide an answer based on behaviuor patterns of existing customers to predict future behaviour, and requires the customer to manifest herself (which usually is an exceptional case rather than a common one), so there is room for incorporating real-time in this.

    Talking about referral value, Lithium Technologies recently published a white paper on Customer Networks and Successful Word-of-Mouth Marketing which tries to put a value to WOM in terms of acceleration and acquisition.

  5. Mark Tamis says:

    Hi Scott,

    You were actually the one together with Mike that got me to look at this question, which I think is an essential step to introducing Social CRM concepts. I too have questions still about its ability to provide a satisfactory customer experience. What I do think it points out is that you need to find the ‘trip wires’ of behaviours and doing an action to ensure the customer does not defect – Customer Experience Management activities can be classified as such actions.

    What were seeing for example in the Airline industry that people would tweet out their discontent, and the playbook could call for taking this to the Direct Message channel. These DMs could ask for reservation number for example so that the link can be made to the customer and that can be used to check with the data in your CRM. From there on you can decide how much effor you should spend on the “case”.

    The online social networks could provide you with additional possibilities for detecting bahvioral patterns, allowing you to react not only to shifts in individual customer’s behaviour but also the overall trends in your customer segments (and even validate that your segmentation corresponds with what you have figured out in the analysis of your CRM data.

    It is but one small part of the overall picture, but a first step for getting a virtuous improvement cycle in place.

    Thank you for sharing your experience!

  6. Jim Novo says:

    Mark, thanks for the kind words about the book, and I think your unwinding of the social side of this above (Enter the Social Customer) is on target.

    If I could make a suggestion, taking a cue from your inside-out / outside-in reference: I think social folks would find more success by not focusing so much on measuring engagement, but instead by measuring *dis-engagement*.

    After all, engagement is the desired outcome; things are good if the visitor / prospect / customer is engaged. Where you being to lose economic value is when the process of dis-engagement begins, and there is some window where action can be taken before loss of customer / reputation occurs.

    Said another way, instead of trying to measure the value created by social, which I’m afraid is fraught with problems, one might more accurately and consistently measure the *value lost* due to the improper handling of social, value which could lead to measuring the value regained through a social CRM / intervention program.

    Here is a very simple micro example. My wife has a friend with small local shop who was looking to see if social could help her business, and I got pulled into the fray. We came up with a simple plan using Facebook / Twitter that appeared to generate business for her, and the most engaged customers would frequently re-tweet her messages.

    Later on, the shop owner noticed several of her best customers had not been in for some time – a tripwire in the classic offline sense. Checking the tweet-stream, we found these customers had stopped re-tweeting shop messages shortly after their last visits. A query to these best customers uncovered a service flaw which was quickly addressed, the customers were made aware of the action taken though a special targeted promotion, and resumed their normal pattern of visits to the shop.

    Now, I realize the above is a long shot away from an enterprise social CRM system, but the fact remains (an opportunity you pointed out above) that the social ecosystem provided a dis-engagement tripwire alert *before* the transactional system. A dis-engagement cycle was kicked off, which then evidenced in the transactional system, where it was recognized (late) and action was taken.

    From the perspective of the shop owner, she can’t really measure the incremental sales value of her Twitter activity or that of her followers, the network effects as it were. In many ways, this social incrementality is simply the value of having loyal customers.

    However, there was tangible measurement of value lost, then tangible measurement of value regained through recognition of and intervention in the dis-engagement cycle. The shop owner didn’t need an LTV model to understand the benefits here, and the average CFO probably would not either 😉

  7. Mike Boysen says:


    Excellent breakdown of Jim’s stuff. And Jim, as expected, you focused right in on the measurement of disengagement, which is so dead on I don’t know why it’s not discussed more. And while a simple example, the use of social data this way is an excellent example of how many *social* implementations are simply monitoring everything and are still in search of the value. Awesome topic.

  8. Mark Tamis says:

    Hi Jim,

    Thank you very much for your comments! 🙂

    Measuring “lost value” through dis-engagement and identifying change triggers in the “social channels” is indeed a topic worth spending time on. To me it also has the advantage in terms of helping define a more manageable strategy for dealing with “The Social Customer”. This will allow businesses to allocate their resources to customer retention more effectively without the having the impression of being forced to drink from the Social Media firehose.

    How much we miss out on when an otherwise loyal customer leaves is an interesting measure to follow and can be put in direct relation to the effort we spend to retain her. The question that bugs me is that in order to measure disengagement “value”, how would you define your baseline to measure against?

    Another point that struck me in your comment is that there is a continuity of behaviour across different channels. Regulars no longer dropped by in the shop and this also manifested itself in the tweetstream when the link was made à postiori. The question to me is whether this also would have worked the other way around – less customer “engagement” in the “social channels” that predicts changes their buying behaviour.

    Being able to predict customer behaviour by trending them cross-channel and finding significant triggers can be an extremely powerful business tool. This in combination with your approach provides a nice framework on which to base sound business decisions that provide value for both the customer and for the company (and not only monetary value). This also has the merit of bringing some rationality to the hype we are currently seeing around some of the discussions around Social Media Marketing…

    Food for thought on how to come up with something useful 🙂

  9. Mark Tamis says:

    Indeed Mike, this may actually be the “actionable insight” that makes business sense rather than the agument of “this is the right thing to do” type of argumentation…

    Someone said (Drucker?) “The real purpose of business is to attract and retain a customer in a cost-effective manner”. I believe that all other business activities are derived from this simple statement. Would you agree?

  10. Jim Novo says:

    I’d think of the “continuity of behaviour” issue like a series of escalations in a customer risk management model, going either way channel-wise.

    I have a high value customer who appears to have defected in the transactional world based on visit patterns. I check social behavior, is it unchanged? Then risk of defection is lowered. If social behavior confirms, then defection risk is increased. Could easily start on social side though, and same kind of “verification” on transactional side could take place.

    Overall, I’m just looking for better action tripwires, and one behavior confirming the other makes a ton of sense to me. I’ve often used service data to do the same confirmation thing, but as is often pointed out, that requires an e-mail or picking up the phone, which may not happen. So social adds some sensitivity you did not have before.

    Not sure I am following your question on baseline, was thinking the value of customer is known so loss could be approximated. Certainly could do controlled testing (in the example, reach out to one customer, not the other) if you wanted a true incremental $$ lift. Sometimes though, the value is pretty self evident and whether it’s $5 or $20 it’s still to the good. % customers active is a good proxy for success without nailing the numbers, if the numbers create too much sweat.

    I take the same general view of LTV – what I want to see is evidence of it rising as opposed to falling. The actual number may not worth arguing about, at least when trying to just get started on doing something tangible.

  11. mikeboysen says:

    Jim, now we are going to have to monitor you for disengagement from this discussion so we don’t lose any value here 🙂

  12. Mark,

    Welcome to Analytical CRM 🙂

    Just kidding, but what you are advocating is actually better use of the existing data to predict future behavior. I believe in that totally, and I think you will find out that most people who have common sense do as well. While I am not a big fan of using value-anything as a metric (it is too subjective, there is no definition and varies from interaction to interaction), I do subscribe to the concept of using data as a measure of success down to the interaction. So much so that I wrote a three-part series on my blog about the best way to do what you advocate (based on my real-world experiences). I don’t like to add links to post, is like spamming and trying to take the glory from the post, so let me know if you are interested in reading it and be glad to send you a link.

    I like you approach, just wish you would lose the “value” label and fuzzy metric. There is a lot you can do without engaging in a discussion of value that will take you nowhere in real life. Else, give me an example of any CFO that has subscribed to any concept based on value (as you say, figures are what they crave, not concepts). Besides, the concept of value is so over-abused by companies trying to prove they deliver it to their customers that is becoming a moot metric in addition to a fuzzy one.

    Nice job.

  13. Mike Boysen says:


    Jim doesn’t really advocate calculating lifetime value. Actually, he focuses on tracking the change in LTV. For instance, building a matrix to track whether older customers are spending more in recent periods could be an indication of increased value if your goal is to retain customers as opposed to continually trying to acquire new customers in current periods. Obviously, that would be a more valuable customer since the cost to acquire is nominal. That doesn’t mean stopping new customer acquisition, BTW.

    So, it can be very simple to build a system of measurement that any business can use to determine if they are making advances in customer lifetime value. If you can’t verify that this is taking place, then you’re wasting your money on CRM – and that includes social CRM.

    Nothing fuzzy about this

  14. Mike,

    Value is a fuzzy concept in spite of whatever you want to call it. I agree with the calculations and the analysis, don’t agree with calling it LTV. Call it lifetime cucumbers and I am 100% sold, it is just the use of the word value that I don’t believe in.

    The rest of the argument, as I thought I said above, no problems.

  15. Sameer Patel says:

    Really good stuff, Mark.
    Focusing on your most high value customers makes a lot of sense. Where I think it gets challenging in the context of the new Social Customer is making that distinction as you get closer to identifying what goes into the top of the funnel. As Rachel Happe expertly opined in a recent post, the top of the funnel is really really wide as you fold in social gestures and in the near term, marketeers will struggle with identifying “high value” (as in, propensity to form a strong relationship and purchase) vs vocal/evangelical/etc.
    I like Jims notion of action trip wires. It sets a more relevant tone to move from a vague definition of value towards a more disciplined approach to identifying propensity to influence and/or convert.

  16. […] this light, I read a recent post by Mark Tamis which he wrote after reading about data-driven marketing in a book by Jim Novo, called Drilling […]

  17. Scarlett says:

    The word is “kudos.” No apostrophe..and BTW adding an apostrophe to words (even those that do legitimately end in a vowel) does not make that work plural; it makes it possessive.

    “Thirty pizzas were delivered.” (plural)

    “The pizza’s crust was soggy.” (possessive)

    Hope that helps.

  18. Mark Tamis says:

    Yes thank you, you’re quite right ( What would really help, however, is that you share your opinion so that we can keep the conversation going… as a non-native speaker of english I am more interested in how my ideas come across as compared to the way I do my spelling. Tell us, what are your views on the subject?

  19. Beth Goldman says:

    This is a great start. If you want to take social CRM and sales to the next level to know when to sell, who to sell to, and what to say, you should check out SalesView by InsideView. You can check out the top 5 reasons you’ll love the new SalesView here:

  20. Mark Tamis says:

    Hi Beth,

    I see that there is an interest in a B2B Sales environment, but how does this help in identifying triggers in consumer buying behaviour? Is there something I overlooked in your solution?

    Thanks for dropping by

  21. […] this light, I read a recent post by Mark Tamis which he wrote after reading about data-driven marketing in a book by Jim Novo, called Drilling […]

  22. […] Found at Data-Driven Social CRM « Social CRM ideas by Mark Tamis. […]

  23. bebenajib says:

    The best thing to hold onto in life is each other.


  24. Vivek Chopra says:

    Read all the posts with increasing interest…I am a huge fan of Jim Novo and feel that his concepts have a huge role to play in the evolution of social CRM…..CRM does have the element of taking cognizance of customer identities, even though it maybe at a very basic level and herein lies the major challenge, with traceability of bloggers to their respective customer records being an issue- even so the whole task of tracing every blogger to his individual records, tracking whatever he posts and making sense of them, seems like an impossible one and there is almost a mental block which would prevent most of us to even embark into such an initiative. However, the current social media monitoring tools do have something on offer which makes the entire exercise more palateable to a normal human mind. Most of these tools have something called “influencer analysis” which identify online entities/authors who seem to have the most influence on the others. If one really meditates on the meaning of influencer, from first principles, it seems like influencers are the customers with maximum impact on my throughput- in fact they can be interpreted to be something akin to the “bottlenecks”, which can influence max number of customers and positively effect my sales if they are happy. So why don’t I start my entire social CRM intiative from influencers? Influencers should not be very large in numbers and so it should be easier to start off by tracking them. So, here is the deal:
    -Identify the influencers as tracked by the social media monitoring tools
    -Find a way to track them to their individual customer records (and this may be difficult in some cases)
    -Invoke Jim’s concepts- form a database of only the influencers and segment them on current and potential value based on Recency, Frequency and Monetary value. Lets call this IRM (Influencer relationship Management) 😉
    – Do a sentiment analysis on only the influencers and track them individually, if possible. Identify the new ones that enter.
    -Monitor if the total LTV of the entire influencers is increasing or decreasing with time (using RFM off-course)
    – What should be the goal of this IRM initiative? To ensure that 100% of the influencers are moved to “high-current value” and “high-potential value” area in the matrix? That can ensure that my company is in good hands..

  25. Jim Novo says:

    Vivek, your suggestion tracks with what I had in mind.

    The influencer cohort could be used as an “early warning system” of trouble to come. If this group exhibited signs of dis-engagement then one might look for the trigger internally – changes in product mix, service policy, competitive landscape, etc. – or could just ask the influencers.

    I’d be willing to bet that in most cases, the influencers would not track 1-to-1 with “best customers”; it would also contain “up and coming” customers and would not contain all best customers – perhaps not even most of them. So it would be important to determine the mix of value and communicate appropriately.

  26. intelestream says:

    Good article. No doubt social networking is redefining how businesses do business. CRM Vendors are developing applications to engage with customers in more efficient ways. Intelestream is about to unveil intelesocial, which integrates LinkedIN, Facebook, Twitter and more with the online CRM, intelecrm. We have also recently published a whitepaper that defines the concept of Social CRM, offers strategies that can help organizations better leverage social networking as part of their overall customer management strategy, and outlines steps that businesses can take to develop a tangible integration between social networking and traditional Customer Relationship Management. The paper can be read at