Bridge The Gap Between Online and Offline

As online complexity continues to grow, marketers are searching for cost-effective ways to reach customers and prospects in channels like social and mobile. Paradoxically, as techniques like segment targeting and content personalisation become increasingly important, they are harder to achieve. This is partly due to the difficulty of applying the kind of consistent audience measurements and targeting techniques familiar in channels like direct mail, TV or telemarketing.

Even five years ago, PC browsers and conventional websites were the hubs of online activity. Now digital innovation has created numerous niches, channels and segments, with mobile and social platforms at the head of these transformative technologies.

With half the UK’s population using social networks (eMarketer 2013), sites like Facebook, LinkedIn and Twitter are vital battlegrounds in the fight for customers – but are also often “walled gardens” where marketing access is tightly controlled.

The burgeoning number of mobile devices running proprietary operating systems like iOS and Android, independent ad networks and a plethora of apps are similarly part of today’s “Splinternet”. By May 2013, 65% of all phones in the UK were smartphones and penetration is still increasing: smartphones made up 85% of phones sold in the previous three months. (Kantar WorldPanel).

Ensuring marketing effectiveness within this increasingly complex world presents many challenges, particularly the need to understand and fully exploit relevant technology whilst managing the spiraling volume of customerrelated data. The consequence is that the big questions are becoming harder to answer: which channels do your customers prefer, where should you spend your marketing budget for greatest effect – and perhaps most importantly; how do you measure ROI?

This white paper discusses how, with help from external data and services, individual-level customer and prospect databases can be extended to bridge the gap between offline and online data-driven marketing and so provide accurate, consistent targeting across multiple channels.

Use What You’ve Got: Why Build A Single Customer View

Perhaps unsurprisingly, instead of leaving customer data scattered around the business in silos, bringing it together in a single database gives marketing the clearest picture of each customer and their attributes; hence the name Single Customer View (SCV). The single view of the customer has long been the key to intelligent datadriven marketing as well as having application right across sales, finance, compliance, product development and beyond.

The primacy of SCVs within the enterprise is now well understood: Gartner’s survey European Organizations Crave a Single View of the Customer positioned adding relevant customer data and information to the SCV as “the biggest challenge in 2013”.

Customer information within an SCV traditionally starts with address and email but can encompass many other indicators and metrics: age, income and occupation, aggregated purchase history, multiple segment codes, net income contribution to the business, propensity scores, web browsing behaviour or relevant market research.

Informing both analytical and operational work, an SCV helps to maximise the value of every customer interaction and communication. An accurate SCV makes it possible to segment customers into homogenous groups, bringing together the most profitable or most loyal – and those who actually cost the company money. Predictive analytics uses historical data from the SCV to forecast customer behaviour. Applications can include customer purchase, lifecycle or risk predictions.

Operationally, that means using the SCV to decide which upsell offer to target to a certain customer segment (and when to send it), to identify who is about to defect (and take pre-emptive action) or selecting those that have already lapsed and making appropriate offers to win them back.

The list of applications is almost endless. Triggering outbound offers based on insurance renewal date, deciding the next best action in real time at the contact centre or recognising offline customers when they register on a website; all depend on the SCV.

Building and Maintaining the SCV

The process of building a SCV is supported by what is known as Customer Data Integration (CDI). This involves cleaning and verifying source data sets, then linking each data item to the correct customer. Data matching – deciding which item belongs to which customer and removing duplicates – is the most challenging part of this process, typically involving comparing names and addresses as well as using unique fields such as email addresses to link to other individual and household-level data.

CDI is a complex business where experience counts: Expert service providers not only have the software tools and skills honed in numerous engagements, but products like Acxiom’s AbiliTec can also reference comprehensive databases to decide which of two conflicting records is in fact correct or to find the most up-to-date individual contact data. That means a more accurate customer database, built more quickly and with far less risk.

Achieving and maintaining this single, integrated view of customers and products is never a “big bang” project. Instead, it is more of a continuing and ever-evolving journey, with constant upkeep critical to preserving SCV accuracy and so effectiveness.

With 7.2 million people moving home and 600,000 dying in the UK alone each year, external databases are a vital part of this refresh process. By referring to these comprehensive national files, it is possible to find new addresses for goneaway customers, update postcodes and perform a host of other mundane, yet vital tasks. Services like Acxiom’s AbiliTec Digital extend the remit of reference databases to online channels, identifying changes across social, mobile and display as well as multiple devices.

Because of the niche skills and investment required, outsourcing all or part of the SCV building, hosting and maintenance process has been marketers’ preferred choice for decades. At one Acxiom client, a large car manufacturer, implementing hosted central customer matching and cleansing reduced redundant processing and saved more than $1 million in annual costs.

This means marketers can concentrate on what they do best – marketing.

Supercharging the SCV

Try as they might, many companies can’t populate their SCV with all the data they need to fully understand, market to and manage their customers. That might be for any number of reasons, from internal barriers and systems issues to the fact that they are simply unable to collect the types of data required.

As an example, FMCG manufacturers have long struggled to collect customer data. Selling indirectly through retailers, they may never actually deal directly with individuals. Though coalition loyalty schemes and pack promotions gather some customer data, those selling indirectly must often rely on other techniques like market research that only give a partial, sampled view of customer behaviour.

This is where external data sets can really make a difference. Segmentations such as Acxiom’s Personicx behavioural classification are built from multiple sources, such as multichannel lifestyle surveys. The individual-level InfoBase file offers these Personicx segmentation codes along with lifestyle, demographic, multichannel and behavioural variables on more than 46 million records representing 90% of UK households.

Added at an individual or household level to an SCV, consumer lifestyle indicators, geodemographic codes and other proprietary segmentation system information offer a ready-touse way to define different customer groups. In the classic application, these profiles can then be used to select the most likely cold prospects for direct channel acquisition.

This is the approach that Unilever took when launching its Persil Small & Mighty liquid detergent. Agency OgilvyOne used Personicx to profile a database of responders to previous Persil marketing campaigns, then selected cold data with similar characteristics. The resulting campaign was a resounding success, with some Personicxselected data cells actually performing better than cells of proven previous responders.

Extending the SCV Online

We’ve already seen how external data can help build an offline SCV and bolster its targeting effectiveness. But today, the SCV must go further, linking to digital channels to support personalisation, offer management and the myriad other challenges of online marketing.

A solid understanding of historical behaviour and the likes and dislikes of each individual is needed to select the correct content and offers for customers. That might mean using cookies to link an anonymous browser to the products they research on a website or logging which emails have been opened, read or clicked.

But a look at Gartner’s “Digital Marketing Transit Map” makes it clear that digital marketing is far more intricate than anything in the offline world. The sheer number of components alone is breathtaking – paid and organic search, social, mobile, ad serving and attribution, analytics and real-time decisioning to name just a few.

With this complexity, it’s not surprising that only a minority of marketers employ existing customer information to inform targeting. An Economist Intelligence Unit (EIU) study found that only 24% of marketers consistently use data for actionable online insight.

External data and services offer a way to apply the offline customer knowledge within an SCV to online targeting. Work might start with understanding how relevant online is to a business: how many and what sort of customers are likely to use online channels and which channels are they likely to use?

Products like Acxiom’s Personicx and the individual-level InfoBase Channel Predictor help answer precisely this kind of question, showing the online usage of different types of UK consumers and their channel preferences across email, post, phone, text message and social networks.

Applied to the SCV, these help indicate the right mix of channels to employ for each consumer segment. Where contact details like email or phone are missing, the same external files can often be drawn on to fill the gaps.

Email: The Critical Link

Of all the channel choices, email has emerged as the most effective in linking the disparate digital world. Versatile, trackable and incredibly low cost, email is the marketing glue that helps hold the Splinternet together.

It can be variously used as an outbound promotional tool, to deliver transactional messages, to manage ongoing relationships, as a way to drive traffic to ecommerce sites or social media – or as a unique identifier that links channelspecific data within the SCV.

Buyers rank email as the most important channel for purchase influence both pre-purchase (37%) and post-purchase (52%), far outstripping social media at 16% (EIU). Yet gathering accurate email addresses and acting on the information that behavioural email campaigns deliver remains underappreciated; the same study found 68% of marketers didn’t rank email expertise as a vital digital marketing skill.

3rd party data and services have huge application here: expert providers can append emails to customer records within an SCV where there were none previously, opening new channels at a fraction of the normal acquisition cost. Customer insight built offline can then be used to inform everything from email treatment to timing, while the email itself can, in turn, act as a link to a host of other external information such as social media ID and enable targeting on social media applications.

Services like Acxiom’s “Email to Banner” are able to retarget email responders via online display campaigns and work hand in hand with a DSP by setting a cookie on the machines of those recipients that have opened or clicked through and associating that cookie to the email open event. The DSP then buys impressions corresponding to users that generated that event by targeting the right cookies.

The targeting of these ads can be refined based on customer data within the client’s SCV. If this is a prospect campaign, segment data from external data sets like InfoBase can be used to target ads accurately and in so doing, drive down costs.

Take Online Targeting Futher

Web targeting is a fast-evolving and complex area of online marketing. Where applied by a company on its own website, there are significant technical and data-related challenges involved. Most proprietary “test and target” website applications are part of larger, expensive software suites and tend to use only single-channel data (typically anonymous historical browsing behaviour), rather than multichannel historical information.

If a multichannel SCV does sit behind a website to support recognition and personalisation, it will require frequent, complex and often bidirectional refreshes. This combination can make building and maintaining in-house systems risky as well as resource-intensive.

In-house-run customer recognition and targeting is also necessarily restricted to a brand and its partners’ own online properties. There is an alternative - one that has been around since the start of the commercial web: online display. This has tremendous reach and scale, with ads delivered across otherwise hard-to-address environments like social media and mobile platforms.

However, with ads traditionally bought separately for each network, planning display campaigns has been an inaccurate and long-winded process with little individual-level targeting. This is changing fast.

Display advertising managed via Demand Side Platforms (DSPs) and therefore leveraging Real Time Bidding (RTB) mechanisms has emerged as a powerful, easily-accessible way to target customers and prospects based on rich individual information. RTB will control a predicted 25% of all online display ad spend by 2015 (eMarketer) in this diverse and complex landscape, depicted in Kawaja’s LUMA Display Advertising Landscape on the following page.

DSPs let planner/buyers centrally manage online unsold inventory available via multiple ad exchanges and networks, with integrated reporting. They are suitable for brand awareness campaigns as well as direct response campaigns: the algorithms that make the automated bidding decisions can optimise the buying towards a specific goal like hitting a maximum cost-peracquisition. Combining techniques like pixel targeting and integration with Data Management Platforms (DMPs), DSPs bring together huge amounts of 1st party behavioural data and 3rd party lifestyle, demographics and behavioural data to enable precise audience targeting via billions of possible impressions.

Targeting options range from page context, immediate browsing behaviour (usually 1st party i.e. generated from the advertiser’s site) to geographical location, demographics and interests. All of this can be used to customise marketing messages, frequency, ad format and much more.

Bringing the SCV into Online Display

Offering a way to reach multiple ad formats and target browsers across many different online properties, display advertising spend grew 26.3% in the first quarter of 2013 alone (Nielsen). But though the audience segments used for ad targeting are often (though not always) detailed and accurate, they are not specific to an advertiser and are difficult to tie back to CRM segments.

This is where individual-level national reference files like InfoBase are able to act as the link between offline knowledge held in the SCV and online targeting variables. Companies like Acxiom leverage their data asset and customer integration technologies to assign a unique and persistent ID to every individual. These technologies, when incorporated into DMPs (Data Management Platforms) and applied to web browsers, enable advertisers to use their CRM data to target selected audiences with relevant display ads. The same approach has long been used offline to link the customer database with media like outdoor advertising, TV and radio.

There are different options to match advertiser CRM data with a web browser. Services like Acxiom’s Collaborative Targeting enable accurate, consistent targeting of customers and prospects, it’s light years away from simply retargeting based on an individual’s previous behaviour.

The advertiser first builds a segmented campaign file of customers and prospects, and selects the ads to be shown to each segment. A neutral third party like Acxiom then matches segments on the campaign file to lists of registered users from publishers. So when a user present on the file, logs into a participating publisher’s website, they will be served with the relevant ad and the response tracked back to the original selection by the advertiser.

With some apprehension over the tracking methods that enable targeted display advertising, this type of solution balances the need for accuracy with privacy concerns. No customer data is ever transferred to the publishing site or combined with other advertiser data.

The process described above only works for direct inventory sold by premium publishers that have a logged in environment. In the RTB world, where impressions get traded one at a time through a number of supply and demand platform integrations, matching happens via cookies. Here, advertisers can utilise DMPs to “onboard” their CRM data and match it to existing cookies.

Targeting Social Media

To date, outbound targeting and acquisition via social media has almost been an oxymoron. Each social network must be dealt with separately and, with little potential for automation, social campaigning can be a costly manual process.

Rather than engage customers directly on social sites, many companies instead prefer to run “drive to social” direct campaigns to consumers via other channels.

External data and services have an important role to play here. Targeting systems, like Facebook Custom Audience offer a simple way to select customer or prospect groups with the highest propensity to post on social media (along with other desirable attributes). Products like Channel Predictor help to choose the most appropriate contact channel while other external data on individual interests, motivations and so forth can be used to inform the tone and content of the message.

Display ads are the other automated option for putting a brand or responsive message in front of a targeted social audience. For example, Facebook’s Custom Audience application lets marketers upload their own customer data and, by matching it (via email address or telephone number) to Facebook’s user database, they can precisely target ads to their customers that are active on the social network.

With Custom Audience, advertisers can anonymously match their audiences to Facebook users and Facebook Power Editor (Facebook’s Ad Centre) allows advertisers to extract a tag that they can place across their site to track conversions and potential sales.

The latest initiative from Facebook is the integration of “Partner Categories” which are sets of 3rd party variables (lifestyle, demographic etc.) from data companies like Acxiom and Epsilon. These sets of variables are available for advertisers to target just as they would conventionally target for cold, outbound mail or email prospects.

Some agencies and data suppliers like Acxiom create their own Facebook agency accounts to run campaigns on behalf of advertisers and, most importantly, apply more bespoke targeting within Custom Audience.

One example of bespoke targeting: The agency first profiles the client’s high-value customers against its national database and then uses those profiles to build target groups of “lookalike” prospects. Those groups are then loaded onto Facebook via the Custom Audience functionality to find matches. Many companies have achieved great results targeting Facebook users with relevant ads in this way lifting click-through rates by up to 40% compared to segmenting using Facebook variables.

Taking It To The Next Level

In this white paper, we’ve looked at how external data and services can help extend the solid, coherent intelligence contained in customer databases to aid targeting within today’s fragmented online channels. It’s also worth remembering that these external databases and segmentation systems can help pureplay e-tailers paint a real-world picture of their customers.

We’ll finish with what looks like the future of the online SCV: the Data Management Platform (DMP). Described by Forrester as the “Audience Intelligence Engine For Interactive Marketers”, 77% of panelists in a late 2012 IAB survey said DMPs will play either a “critical” or “major supporting” role in expanding the performance of long-term advertising and marketing efforts.

DMPs today are typically employed by advertisers to support the complex business of targeting within the RTB environment. Some DMPs partner with big publishers, packaging the visitor populations into actionable segments that can be targeted with display ads.

However, consumer data companies like Experian and Epsilon are increasingly partnering with DMPs to enhance online audience intelligence, effectively enabling the activation of offline 1st or 3rd party SCV segments in the digital world.

As well as helping publishers deliver segmented audiences to advertisers, DMPs support enhanced web visitor recognition and, depending on the other individual-level data they hold, can drive improved offer management, website experience optimisation, reporting and a wide range of other functions. Essentially, the DMP lets marketers segment their audience and then project that across multiple interactive channels.

Just like the SCV, some companies will choose to build and maintain their own DMPs while others will outsource. With vendors like Acxiom offering reliable audience intelligence across online and offline channels, they are well placed to provide a solid foundation for a consistent, accurate DMP and according to Forrester1 are “investing in the technology and publisher partnerships necessary to stitch together a massive pool of targetable consumers across platforms”. Acxiom have taken it to the next level with the launch of their new Audience Operating System (AOS), “a breakthrough in multichannel marketing” says Gartner analysts2. The AOS will allow marketers to consolidate the clutter and complexity of data across media, channel and devices with an ever increasing level of effectiveness.

The single customer view remains at the heart of intelligent marketing, but marketers must strive to extend its remit to the complex and challenging online world. Continually balancing the benefits of expert external data and services with investment in relevant internal skills and technologies is the only way to achieve this.