How To Use Customer Data For Accurate Cross-Channel Targeting

by Michael Kaushansky, Tuesday, July 23, 2013 12:48 PM

We are quickly moving to a world of cross-channel hyper-targeting.  The notion of database-precise targeting is not just for direct response campaigns anymore, but now can be used for brand campaigns.  The promise is to provide a laser-focused effort in delivering not just an offer, but also a brand experience to a targeted individual.  The most critical piece in the process is to have the right data to know whom to target.

For most brands, their existing customer data is the most effective method of defining which future custo

mers they should go after.  Sounds like a simple task, yet in reality it’s anything but.  Several barriers exist, including the low accuracy of “cookie-synching,” and misclassification of customers into the wrong segments by relying only on online activity. Most importantly, some brands are simply not set up for access to their customer databases.  Brands must store their customer data in a simple and accessible manner, then carefully select the most reliable data sources/providers and enhance each customer’s profile.

Even with properly organized customer data, some advertisers have reported offline-to-online match rates as low as 10%, with only 50% accuracy in simply identifying gender, let alone deeper attributes about a consumer. 

To bring consistency and accuracy to data matching, analysis, and targeting, I have defined a five step process: AESMA.  These steps help convert valuable/proprietary first-party information into a highly competitive consumer targeting strategy:

1.     Access – Enable selective access to customer data, ideally using a data warehouse.  To improve match rates and data accuracy name and address as essential, email helps but is not required. 

2.     Enhance – Select a data enhancement provider with proven match rates above 50%.  There are a handful of primary data providers who aggregate consumer data using name and address for matching.  The match-key is kept fresh using the NCOA process (USPS’ National Change of Address).  Secondary data providers rely on these primary sources, you may as well go direct.

3.     Segment – Develop a dual segmentation framework using data native to a brand to create intuitive/proprietary segments and use third-party data to create empirical/data-driven segments.  A matrix between the two segmentations will provide a deeper understanding of existing customers and help define a strategy of whom to target in the future.

4.     Model – In most cases you will end up with too much data.  Use modeling techniques to isolate and remove redundant data and highlight only the handful of variables that matter in identifying your target prospects.  For example using Household Income would reduce the need to also use a wealth indicator, etc.

5.     Action – Having all the data and analysis is only half the story unless you do something with it.  Use the modeling output to help craft a more focused campaign across all channels.  Certainly digital channels allow for more targeting parameters (e.g. DSPs, Facebook, and email); howeverm you can also aggregate results by Zip code and target using traditional channels (OOH, Television and radio).

Soon all channels will allow for some level of hyper-targeting, even in television with the advent of set-top-box targeting.  Starting to enhance and segment your customer data now for CCHT will provide enough background and experience to be ahead of the curve.

2 comments on "How To Use Customer Data For Accurate Cross-Channel Targeting". 

Grant Bergman from SurveyConcierge.com • GrantBergman.com commented on: July 23, 2013 at 1:45 p.m.

I think you need a preparatory step to "Identify" or "Inventory" data assets for marketing use. In my experience, MANY small to mid-sized companies (even a couple $100 million in revenues) do not have analytiucally-attuned Marketing leadership that knows what information is available to them. Yes, with the hyper-focus on online behavior they all know they have Google Analytics to play with. But what about actual customer files that tell us about the whole customer and not just their online behavior once they enter our funnel? More than once I have started a consulting engagement asking to speak with a CFO or Controller: All those customer files that are necessary for billing and delivery of services contain real marketing intelligence about the business's clientele. Simple geographical profiling (we have all the customers' ZIP CODES, right?) can help us back into demographics that help us to find "look-alikes." A look at repeat business can offer insights on lifetime customer value. If the business follows a subscription model, the same data tell us about retention rates, both now and over time. Line these up with promotional efforts and we can start to diagnose the ROI behind marketing spending. And since we already know all our customers' email addresses (we do know this in 2013, don't we?), outreach to get more input is as simple as a cheap online survey (crafted, one hopes, by someone who knows something about developing, deploying and reading surveys). Or we can pick off a handful of highly involved customers, based on their purchase history, to do outreach via qualitative research, social media, or advisory panels. Note that none of this requires building a data warehouse infrastructure, at least not at first. The main issue is knowing what data are available and who can help us to access them.

Paula Lynn from Who Else Unlimited commented on: July 23, 2013 at 2:34 p.m.

You mean hyper spying.

Read more: http://www.mediapost.com/publications/article/205131/how-to-use-customer-data-for-accurate-cross-channe.html?print#ixzz2ZtZeMBzH

Share this post