The evolving data architecture of social customer relationship management
- 1.3k Downloads
Social media enables the brand to extend its personality to engage with consumers on their terms, at the time, place and channel of their choice. Replacing traditional customer relationship management (CRM) with a Facebook fan site (for instance) presents advantages but also strategic risks for a brand. If knowledge of brand advocates, High-Value Customers, High-Value Prospects and influencers is considered an asset for the business, the asset needs to be owned and managed over time, and that is not possible in the social-only environment. Social and CRM in combination are likely to be more effective than individually. In linking social data with CRM data, multiple identities, privacy and data protection issues, and which data to hold on the database (vs. leave in the social media ‘cloud’) are all challenges that need to be addressed. The social customer relationship management (SCRM) strategy will be driven by marketing, but will have an impact through all the business functions from customer service through to brand marketing. Eight stages of the SCRM strategy are outlined from ‘Listen and Learn’ through to ‘Measure and Evaluate’
Keywordssocial CRM social media CRM customer relationship management consumer engagement customer database
How ‘Social’ relates to ‘CRM’?
What data is required to support SCRM?
Which of this data need to be stored on the database?
Next steps for brands new to SCRM?
How ‘social’ relates to ‘CRM’
In data-rich markets, such as financial services, automotive, telecoms and some retail, direct marketers have for years combined sales data, profile data provided by customers and varying quality of data from external sources, such as neighbourhood, credit or lifestyle databases. They segment their customers based on value, needs and (occasionally) psychographics to predict the propensity to purchase. They then use sophisticated targeting (outbound to the customer, or inbound when the customer calls in or uses the website) to communicate the right offer to the right person at the right time. These techniques can uplift response rate versus random selections of contacts, but they are still blunt instruments. The interactions are designed to sell, not to engage or build relationships. Although ‘traditional’ CRM helps manage Customer Interactions on a large scale, it does not help in building relationships and mutual trust between consumers and brands. To build relationships and trust, it is necessary to know the customer well and not just be limited to mere ‘transactions’ as was the case with ‘traditional’ CRM.
Enter social media
Pioneer brand marketers are building listening posts and workflow processes to identify and then do something with this insight (eg push sentiment insight quickly into campaign teams) and to decide whether, and how, to react to comments. They are identifying influencers in their market and at least extracting the names and digital personas of these influencers so that they can listen more deeply to what they say. Brand marketers are pioneering the use of social geo-location data (eg via Foursquare, Gowalla, Facebook Places) to understand the movement of their target market (‘check-ins’ at airports, hotels, corporate headquarters tell you something about who you are dealing with) and push local offers to consumers in a certain area (eg Pepsi's ‘Pepsi Loot’; Dominoes pizza — driving footfall through FourSquare). They are beginning to work more closely with retailers, sharing data with each other, driving footfall for the retailer, creating a win-win-win between consumer, retailer and brand.
Social Media and CRM: 2+2=5
Budweiser: Not everyone agrees with a HVC strategy. At the WFA/RVD Global Advertiser Conference in April 2010, Istanbul, Chris Bruggraeve, CMO of Anheuser-Busch inBev pointed out that heavy drinkers do not make much impact on the firm's bottom line. Instead, promoting temperance pays for itself over time by building long-term brand reputation and trust. ‘We are not interested in the 0.001 per cent of people who binge-drink,’ he added. ‘We don’t need them. Half of the world's population hasn’t even tried to drink beer, so if we can get 5 per cent of those to drink one beer everyday, we will be very happy’.
Web strategist Jeremiah Owyang 7 says: ‘Traditional CRM suppliers, such as Salesforce.com and SAP are starting to integrate data from Facebook, Twitter and other online social networks. Dell and Comcast, both leaders in social marketing and support, have already integrated Twitter data to allow brand managers and support teams to actively track what′s being said in tweets’.
What data is required to support SCRM
The objectives of SCRM need to be clear in order to define the conceptual data architecture. Where does relevant data exist, and how can the data ‘cloud’ be managed to achieve the objectives? SCRM objectives must be one or more of the following: to (1) identify, (2) engage with and then (3) sell to people within a target market through building a stronger relationship. These overall objectives are not any different to those of any CRM programme. The difference is having access to data and tools that allow more effectiveness in all of these areas. For instance, the target market could be identified, engaged and sold to in a better way through (1) listening to social conversations and accessing deeper insights into behaviours of the whole market to better engage with them through any media; (2) attracting fans through social content, fan pages and websites; (3) panning social media and fan sites to identify brand advocates, influencers and HVCs and HVPs with a view to developing a greater insight on them, and maybe to interact with them more personally; (4) interacting with (and promoting messages to) consumers when they are most receptive and (5) making it easier for them to buy from us when and wherever they want. Thus, traditional 1:1 CRM data models are disrupted and need to change.
Traditional CRM model
What are the changes in data content enabled or forced by the emergence of Social Media?
Which data need to be stored in-house versus leaving in the Cloud and using in situ?
Data to identify
all the data the user opens to ‘Everyone’ plus Facebook's basic data and profile information can be linked into;
the user's ‘Friends’ provide networks of potentially like-minded individuals.
The social cloud is a hunting ground for marketers. Internal policies and local legislation must determine a firm's strategy around recruitment of names from social media, but with the right permissions brands can filter off these individuals in order to keep closer to them.
Data to match
- 1First, there are ever-increasing numbers of potential records for every physical individual. An avid social media user can easily have 10 different identities in social containers (see Figure 3), covering their personal, social and business lives. In addition, matching criteria typically involves data that are prone to change such as a user-selected name, ‘handle’, email id and mobile phone number (Figure 4).
Secondly, the individual may not want different versions of their existence to be joined together. It would be a brave marketer who joined ‘@smithj12’ on LinkedIn to ‘thebigcool’ on Facebook (without seeking and obtaining the individual's approval) even if they did share the same email address.
Forename and Surname
Date of Birth
Social media ‘handle’ or ID
The latter two are sometimes all a company has in SCRM. It should be noted that in some markets (eg Africa), it is common for one person to have multiple mobile numbers either on discrete devices or in some cases a selection of SIM cards.
On a social site there may be a name or nick-name, handle, email address and mobile number, none of which are capable of ratification and are less certain in terms of matching. While matching is key, it will require robust processes and a clear policy about will be done with the consumer record once built.
Data to understand
Historically, understanding of customers has been based on geo-demographics, transactions, research and possibly a record of other behaviours such as complaints and response history. These will remain important but provide a limited customer ‘understanding’. By monitoring, assimilating and understanding the attitudes and behaviour of individuals in their social environments, it is possible to have a much greater depth of understanding of their likes and dislikes, their thinking, their motivations and for some, their ‘passions’. This will be derived from their self-motivated inputs, not from often misleading self-completed questionnaires. 9 Listening can be database-independent, with software tools rapidly sifting through terabytes of data to identify conversation themes; velocity and ferocity of conversations and brand mentions.
Companies may need to derive nuggets of understanding themselves by taking feeds from the application programming interfaces (API) of social media owners (eg Facebook Connect) and doing the analysis themselves in the short term. However, there is clear evidence that the social network owners see this as a value-added opportunity. Emerging companies such as Flowtown 10 and Rapleaf 11 aggregate and rent consumer data from many social media networks against a consumer record based on matching with data as simple as email ID.
Data to do business
easier for consumers to find and order their products, wherever the consumer is, whenever they want;
easier for brands to link sales to consumers.
Some examples of these are photo-enabled product matching (eg Amazon, AutoTrader), smart phone cameras, intelligent barcodes and 1-click ordering. Mobile devices will enable consumers to download e-currency (eg e-coupons, loyalty points) and purchase items without a credit card. However, from a data viewpoint, this will be the point when a virtual ‘persona’ can be joined to their physical persona's record. This is because the world of real transactions is still based almost entirely in the physical world, and thus the data required to do business will include data such as credit card details (or online payments systems such as PayPal), verified by address that can only relate to a physical persona.
Data to communicate
Outbound CRM communications were sometimes made based on selecting lists of customers with a high propensity to respond, whereas SCRM communications can be real-time based on a ‘signal’ or ‘event’ from a social media transaction. This may be a known customer's comment on a community site, a product review by them, a blog or a ‘check-in’ update. Most of the time, to most of the market, the communication can stay in the social cloud and never touch the database.
Data to personalize
Personalization is less important in the world of Social than it is in the world of CRM. ‘Outbound’ communication via social media is largely about targeting relevant content to the social ‘containers’ that are identified as influential to brand users. One-to-few targeting will be used to respond to comments on fan pages. It is difficult to personalize in the social cloud. Personalized, one-to-one messaging is important over both digital and physical media — on the brand web/mobile site (eg welcoming and messaging individuals with their personalized content), as well as through direct outbound communications to consumers stored on the database (probably high value, brand advocates and influencers). Some website personalization can be achieved without a database. For instance, if a consumer accesses a website, the use of ‘cookies’ can help personalize the experience for the consumer the next time they come onto the site. Some social media containers (eg Facebook) provide profile information if a fan clicks from Facebook, which allows the brand owner of the website to personalize to a limited degree. However, the big opportunity for increased personalization comes from joining together information known about individuals and their social media personas. This would enable relevant content to be pushed to their mobile device, perhaps based on their value and profile, when they check-in at a location, or when they downloaded an application, or played an interactive game or made a comment on a blog. This is emerging marketing territory, but intuitively will enable extraordinarily relevant relationships between brands and consumers.
Data to predict
Propensity modelling has become very sophisticated over the last decade, and many organizations can predict key behaviours (eg next flight, next car service, or policy cancellation) with high probability. The impact of social networks in the area of prediction will be to enhance these predictive models. The information available on the social media containers will enable a better understanding of the behaviours and attitudes of the individual, and therefore a richer ‘propensity to respond or purchase’ modelling is possible. It is possible that detailed data available from an individual's social networking activity can make a prediction of key behaviours even more accurate. A stream of negative comments on Facebook or Twitter about a service experience in a retailer or car dealer, for instance, can help predict a future ‘defector’.
Data to engage
In addition, getting to know what they tend to like and dislike will enable an understanding of the types of stimuli that will encourage or deepen their engagement. Do they respond to news, sports, celebrity, fun or environmental content, for instance?
Which of these data need to be stored on the database?
The data that need to be stored and maintained in a company's database to understand, track and interact may actually reduce with the advent of social media.
identification and matching data;
derived segment data for macro personalization (eg HVC flag, Engagement flag, Influencer ranking, lifestage);
any actual transaction history (eg sales, responses, complaints).
Where should data be held?
Definitely hold on database for valuable consumer groups
Potentially hold on DB
Leave in cloud
Traditional CRM contact and transactional data
Profile data enhanced with codified data extracted from web or social sites
Transactional data either extracted from interactions and stored on DB OR left in cloud and accessed tactically for analysis or promotions
Transient data available ‘in the cloud’ for instant promotions, messaging and/or listening/ analysis
Full contact details
Email ids & Mobile numbers
Website/social container visits
Content of conversations
Socio-geodemographic profiling data
Influencer flag and score (algorithm)
Interaction level (eg via comments; log-ons; games played)
Outbound campaigns, Inbound responses
HVC flag and score (by brand if appropriate)
Comments left by consumer which you chose to respond to
Macro travel patterns; (eg London, Paris, New York)
Like, dislikes — for analysis of content engagement current geographic location (for relevant promotions and to drive retail footfall
Loyalty ‘level’ (eg points)
Locations type (Airports, Hotels, Business centres)
Passions (hobbies derived from categories of on-line activity) (eg music, sport)
Macro segment (eg attitudinal, behavioural) Another layer of social permissions
Relevant Applications downloaded/frequency of use
Social network active membership.
Next steps for brands new to SCRM
- 1Listen and learn: Where are the customers playing in the social cloud and what conversations are they having? The first and easiest thing to do is to listen to the buzz about the brand, product category and customer segment.
- 2Understand consumers’ socialgraphics 13 : A clear understanding of how consumers use social media is key to this and will ensure the strategy is based on real data and not just gut feel or latest marketing trends.
- 3Define SCRM role in customer engagement Strategy: At the enterprise level, brands will need to prioritize the development of a social business strategy that contains within it social marketing, but will fail unless based on strategic business and marketing objectives. These objectives must be to do with identifying, engaging with and selling to customers. Two key policies to support this will be defining whether to and how to recruit customers from the social cloud, and how to identify HVCs and influencers.
Develop data model: Once clear on objectives, the next priority is the development of the data model itself. This should be based on identifying and then acquiring the most relevant structured and unstructured data to allow the delivery of real-time customized website experiences, as well as personalized outbound communications via email, mobile and social channels. It will include rules for codifying detailed data.
- 5Optimize website: Brands should prioritize the social optimization of the brand website, integrating social functionality to turn the site into a dialogue-driven, interactive and engaging experience. At the simplest level, brands should encourage ‘sharing’ and ‘liking’ capabilities within their sites, letting consumers take the brand messages out into their networks of consumers. The Edelman Trust Barometer 14 shows that advocacy, conversations with friends and peers are twice as credible a source of information as advertising. Brands should also offer consumers the chance to log-in and register using tools such as Facebook Connect or Open ID. Brands should also consider the creation of ‘owned’ communities within their own sites (of the type offered by vendors such as Jive and Lithium) to encourage consumer-to-consumer dialogue, customer service management and of course recruitment opportunities.
- 6Recruit fans: The next brand priority should be the recruitment of brand fans and followers within branded social pages. Social networks ultimately represent a gathering of interested and sometimes passionate brand fans, and brands should take the chance to offer a compelling proposition to recruit consumers into managed RM programmes. This can now be simply delivered within a Facebook page tab using a simple API mechanic.
- 7Welcome and activate: All fans, and consumers who have registered, should be welcomed to the brand in a way that is appropriate to the mechanism through which they have become a fan. Activation will apply to those who have downloaded an application and not used it, or have joined a loyalty scheme and become dormant. Many fans and registered customers lie dormant after the first transaction. Mechanisms should be designed to encourage interaction.
- 8Measure and evaluate: The final priority should be measurement and evaluation. Brands should establish metrics and a managed approach to report on the effectiveness and efficiency of their activities based on the objectives set at the strategic level. Many commentators have said that assigning a value to a Facebook fan is difficult and often misleading as it is not the fan itself that has value but rather what the brand does with the fans to drive incremental value. Using standard social metrics, such as the volume of fans, likes, followers and comments, is not a good measure of awareness or brand engagement. Forrester suggests 15 that a balanced scorecard is needed to evaluate the wider picture of the impact of the SCRM efforts, a focus not just on financial measures such as Return on Investment (ROI), but more on Return on Objectives (ROO). These might focus on Financial returns (has revenue or profits increased as a result of SCRM) but also on Digital returns (has the brand enhanced its owned and earned assets), Brand Returns (have consumer attitudes improved) or Risk Management Returns (is the company better prepared to respond to PR or reputation challenges).
Social and CRM in combination is likely to be more effective that either social or CRM individually. Social can be used to engage consumers interested in the brand with content placed where they play and work, through PCs, mobile devices, screens, kiosks, tables and a number of new ‘channels’. Brands can interact with their consumers directly through social media without ever having to know who they are. However, social media containers are risky for brands. They do not own the ‘fan’ data; their consumers can drift to another social container very quickly, and targeting different types of consumers is not easy. In addition, the disproportional importance of high-value consumers, prospects and influencers means that they are strategically important to a brand. Like any asset, this implies that these groups at least need to be managed carefully over time (the term ‘managed’ is used lightly in this social media context). This is done because it protects revenues and develops profit, in a sustainable way, through engaging HVCs. The HVC may be happy for a company to interact with them through one-to-one communications taking the relationship outside of the social container.
Ways brands add value with HVCs?
To identify, match, understand, predict, communicate, engage and sell to these people, requires adopting and refining CRM approaches and the databases that support them.
References and Notes
- Consumer sales data and addressable data.Google Scholar
- Engagement is the emotional engagement with a brand, product or service. From an FMCG angle, Engagement involves participation, interaction and involvement. It is emotional; it is about actions not just thinking differently, but doing something either within a brand experience or after because of it.Google Scholar
- ‘Gap scraps new logo after online outcry’ http://www.bbc.co.uk/news/business-11520930.
- For a discussion of the model, please contact Neil Woodcock from The Customer Framework at email@example.com.Google Scholar
- For a definition of Social Cloud see, Pezzi, R. Information technology tools for a transition economy, (2009), http://www.socialcloud.net/papers/ITtools.pdf, p. 3.
- Steve Ruben, Edleman re potential data privacy 9/11 http://www.steverubel.com/a-privacy-9-11-could-derail-social?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+steverubel+%28The+Steve+Rubel+Stream%29&utm_content=Google+Reader.
- See State of the Nation IV 2005, QCi chapter 3?Google Scholar
- ‘Social Graphics: a term coined “social graphics” by Jeremiah Owyang of the Altimeter Group’.Google Scholar
- The ROI of Social Media Marketing, Augie Ray, Forrester Research Inc., July 2010.Google Scholar