FREE DM Review Site Registration!
Sign-up today and access DM Review on the Web!

Your FREE registration entitles you to:
FREE email newsletters

FREE access to all DM Review content

FREE access to web seminars, resource portals, our white paper library and more!

   

Data-Driven Marketing: What Might Happen Next?

The heightened interest in and use of data to drive marketing decisions is a widely discussed topic. This interest also means that people are working on how to improve current approaches to develop new best practices. Some important marketing topics and trends are data related, some are technology related; but they are all directed at improving the efficiency and effectiveness of marketing activities.

Marketing mix modeling merges with database marketing. The former discipline allows me to allocate budgets to various spending categories such as Internet, TV or direct mail, while the latter lets me rank customers or prospects based on their likelihood to exhibit a desired behavior. When working with these tools in isolation, one is left with some unanswered questions. In marketing mix projects, the CMO might ask: on exactly who should I spend the budgeted amount? In database marketing projects, the CMO might ask: how deep into the list should I go before other tactics are more effective in satisfying our overall objectives?

Wouldn't it be valuable if we linked the two disciplines into an integrated solution? Imagine the ROI of tying optimized allocation with optimized targeting. Because the total marketing budget is reasonably finite, it appears to be an allocation problem. With most forms of the media mix offering discrete targets (or firms switching to those tactics), the link between tool and audience is becoming easier. Campaign management seems to be the application that straddles both areas because campaigns consist of both budgets and targets. Marketing Mix Modeling often indicates when to spend funds as well as on what tactics to use, so it shouldn't be a giant stretch to have an additional output that represents a series of campaigns that feed common tools. On the other hand, the database marketing perspective requires some guidance on how many customers to contact. Marketing mix modeling could fill in the budget amount as opposed to the historic case of typing some number in spreadsheet cell. Because campaigns cost real money, it seems that equating the sum of the parts with the whole would fit well into budgeting and resource management applications.

Steps are being taking along this line, particularly among firms providing value-added services on top of large compiled consumer data sources. The high fixed costs of building a syndicated data source and the perception of third-party data being somewhat of a commodity constantly raises the bar in terms of increasing the value of the solutions offered. So, I'd expect to see services where media allocation problems include a list of targets for a series of campaigns. While this approach was discussed in terms of linking people to the mix, the logic applies equally well to stores or other discrete units where money is spent and results tracked. So, stores should be ranked in terms of their propensity to generate results for trade promotions and so forth.

Surveys aren't just about samples any more. The good news about surveys - I can ask people about any topic I'd like, such as product usage, attitudes and media preferences. The bad news is that the costs prevent me from asking everybody. The result is that I know a lot from a few but usually lack the ability to apply the findings to many. In the end I often have a research report that isn't actionable or must take a stab in the dark at finding more people like those of interest based on gross generalizations.

Wouldn't it be useful if we could map survey respondents to either our total customer base or a large prospect pool? Imagine being able to apply brand usage and attitudes in the selection process of prospects. In fact, the approach used to sample respondents for political and media research as well as omnibus surveys and panels allows for this leap to be made. The practice of stratified sampling tells us a lot about each respondent selected. Of particular interest in this idea is the fact that people live somewhere and it is likely that other similar people are located nearby. By using geocoding, the tagging of location-specific information, a mapping can be made from one group to another. This approach can be implemented with people or even computers by using their IP address for ad serving networks. Employing similar analytic approaches to allocation and propensity problems, it is possible to determine the likelihood that a customer or prospect looks like a member of a survey panel.

With a little bit of ingenuity and some creativity, it is possible to find likely prospects in the total population based on brand usage from a sample. While by no means perfect, it is definitely a step forward from the "target the 18-49 year old market." An obvious extension would be to look at softer items, such as satisfaction and willingness to recommend, both known to impact brand equity and company value.

Marketing finally gets fly-by-wire capability. In a time when responsiveness to market events is becoming more compressed and critical, the typical loop approach to CRM still leaves too much time for changes to occur that invalidate many possible actions. This is particularly true in the online world where customer expectations are radically different than for other channels.

Wouldn't it be more efficient if we could directly influence our marketing activities from the tools we use to monitor their effect? The large and growing investment in business intelligence (BI) provides us with a lot of information but leaves us wondering just how to take action. It reminds me of a cul de sac or dead-end street. The reasons for showing key metrics on a dashboard are correct, but the analogy of being like a car or plane is partially inadequate. Imagine a pilot looking at an altimeter and not being able to do anything about it. As an industry, we're moving from flying blind to flying by the seat of our pants, with the next logical step dealing with direct implementation of decisions, or fly by wire.

The logical starting point for marketing would be in the identification of areas where value can be increased within the business. There are two areas where this is happening. First, the concept of operational BI merits its own discussion and products with the aim of getting information into the hands of those who can take direct action. Second, the people concerned about process are also working on this issue. Understanding business processes should routinely include an assessment of how best to monitor performance. A common thread is that campaign management could be extended to cover these points. Imagine being able to use a marketing dashboard as a selection interface.

Measuring customer behavior, not company behavior. Because marketing funds are spent to alter behavior, why don't we measure and report on individual performance metrics more often? Because marketing is best defined as looking at the "business from the customer's point of view" (Peter Drucker) why don't we use their metrics instead of our internal metrics? Is it because internal numbers are easier to find, or it's easier to justify budgets with them? Is there a philosophical reluctance to measure because we don't actually know what to expect?

Wouldn't it be more constructive to report on individual key performance indicators (iKPIs) rather than just aggregate ones? This way we could measure the direct impact of marketing activities on customer behavior. For each campaign, the objective should be stated in terms of the specific behavior we want to affect. Terms such as share of wallet, purchase/open cycle, dwell time and a host of new ones should augment traditional metrics like response or open rate, cost per activity and other management terms.

With few exceptions, company revenue comes from customer transactions, and the firm's market capitalization derives from the expected lifetime value of customers and prospects. This suggests the firms that understand the direct financial impact of customer behavior have a distinct advantage. In this context, behavior modification is not a four-letter word. An interesting test is to review a number of recorded earnings reports for references to the impact of customers on the business. A common point is, "Sustained growth is function of strengthening our client development process." I'd like to see financial conversations adding a "customer value per share" metric.

The four scenarios portrayed above don't represent wishful thinking or a fancy crystal ball, but some of the work that leading companies are doing to better understand and target customers.


Anthony Power is a principal with Power Consulting, where he is responsible for helping clients create strategies, positioning and solutions for innovative marketing products and services. Power has more than 25 years of experience in the design and implementation of information-based solutions across a variety of industries that lead to market penetration and revenue gains. His experience covers the full spectrum of customer information and technology - market research, systems integration and software development. He can be reached at Anthony.Power@comcast.net .

 

For more information on related topics, visit the following channels:



Industry Vendors