Despite what marketers may want to believe, most who visit their Web sites are destined to have only a brief stay there.
The typical surfer drops into a brand marketing Web site for one reason or another, and then just as quickly goes on his or her way. The marketer is often left longing to start a more in-depth online relationship—which will probably never materialize.
Marketers can lament this hard fact of the online world, or they can seek gains from it. Just because a surfer's visit is short and fleeting doesn't mean that the relationship is without significant value. With a proper messaging approach, a marketer can shape even the briefest of site visits into a gainful interaction.
One such proper approach is to be continually agile with our messaging to short-term visitors through an iterative process of learning what works best at any given moment to create gains with these users, and applying these insights to future rounds of messaging.
The idea is to keep refining a site's messaging in a rigorous direct marketing fashion to squeeze the most gain possible out of even the shortest of contacts with a consumer on a site.
It is a simple idea. This trouble is that most brand Web sites are just not well designed to do this type of gain prospecting.
Brand sites tend to serve best the needs of more involved users while neglecting the needs of transient ones. Marketers are all too frequently just preaching to their choir through their Web sites rather than also seeking ways to influence the behaviors of shorter-term visitors who often are in critical states of consideration or early awareness about a brand.
Additionally, most of these sites lack an adequate technical infrastructure to carry out truly gainful messaging optimization.
Information architecture has done a good job of cleaning up site designs to better serve the navigational and experiential needs of short-term visitors, but this is only a start. Information architecture can improve the visitors' experience of a site, but it does not necessarily provide marketing gains in the process. It also assumes that these visitors' intent is correctly understood a priori and that this intent is unchanging over time.
The success of most Web site marketing messages is left up to the instincts of their creators, which in my experience can be a hit-or-miss proposition. In the same way, the fate of the short-term visitor gains is left to the vagaries of the creative process and the risks embodied in it when there is no significant performance feedback to direct its course.
Marketers should aspire to do better.
Short-Term Visitor Gains
An opportunity exists to be more gainful vis-à-vis our short-term site visitors by embedding dynamic features into our Web pages that first test a set of messaging to understand what works best to garner marketing gains in our interactions with these elusive visitors. Then, these same dynamic features can apply the results of these tests to preferentially show the most gainful messages to future short-term visitors.
With a standard application server and a simple decision algorithm, this workflow can be fully automated without much trouble. Automation will allow a site to adapt in a gainful manner without necessarily requiring labor-intensive intervention by a site's owner.
This concept is hardly new. Automated decision frameworks already exist in interactive marketing, but they have largely been restricted to ad serving and email broadcasting contexts. This is unfortunate. An automated messaging optimization approach could be readily deployed within the typical technical architecture supporting most brand Web sites, with limited effort and much code reuse potential.
The benefits of this type of approach to the marketer are strong and should not be underestimated. Such optimizations, when correctly deployed, will achieve the following:
- Improve the overall marketing impact of a Web site, especially with less involved visitors
- Lower the risk of using ineffective messaging on a Web site
- Provide messaging insights that can be deployed elsewhere in the marketing mix
- Allow messaging to respond to changes in consumer behavior as they happen
The most significant barrier against adopting this approach for most brand Web sites is a mass-marketing-derived perception that Web sites need to be managed like hardcopy magazine content rather than through direct marketing style optimization. If a marketer can get past this intellectual hurdle, the path to improved gains is not terribly difficult.
Messaging optimization work can be done manually. But typically the manual approach has limitations in practice, for these reasons:
- There is a delay between the start of any manual test and its application.
- Human nature often conspires to prevent the marketer from applying the insights of a trial in a timely manner.
- There are commonly only weak safeguards in place to recognize and respond to post-test change if it should occur.
- It is easy to become overwhelmed by the manual management of all the messaging locations that could be gainfully optimized on a standard Web site.
I have previously written here on the importance of flexible practices for success dealing with integrated marketing (“Striking at the Heels of Integrated Marketing”). The reader should understand this article to be an applied extension to that strategic work. The use of messaging optimization tools fits into a set of practices that, collectively, I would call agile or maneuver marketing—a marketing approach that espouses tactical flexibility as a means of providing heightened competitiveness.
Beyond its immediate use seeking local gains, the power of automated messaging optimization approach lies most profoundly in its ability to react to change competitively without necessarily requiring hands-on intervention by the marketer.
This allows for the creation of messaging strategies on a Web site that are highly change tolerant and aggressive in seeking continual gains—a core concept of agility. It also frees up the marketer to deal with other issues and opportunities in the overall marketing mix while remaining gainful with his or her messaging.
An agile approach works in large part by managing risk through hedging the bets that all marketers make when fielding a set of marketing messages. Despite even the most diligent preparation, the marketer will not truly know the effectiveness of a given messaging approach until it is used in a real-world context. Agilely applying performance feedback to tune the messaging executions in the field diminishes the probability of overall program failure from poorly-performing messaging; chances are that at least one execution of a set of messages tested in the field will work, and the poorest performing messages will soon be removed from use.
If one could perfectly predict which creative execution would be most effective with visitors, the use of any optimization methods would be inefficient. After all, the process of optimization requires that a series of perspective executions of varying performance be tested against one another before the most gainful ones can be recognized. Accordingly, an optimization solution cannot be a perfect maximizing solution, because it will never achieve maximum potential performance unless all units perform equally well as one another.
However, at any moment, optimization will seek gains beyond the average gains created by a random selection of perspective executions. In this way, a long shot at a perfect maximization of efforts is traded for the assurance of better-than-average (often, much better) gains overall, which is a great deal in our imperfect and competitive world.
To work, messaging optimization requires defining a currency of value against which the success of the different messaging execution will be measured. For sites that sell directly, this value can be easily understood in terms of how much a given messaging execution shown to consumers translates into increased sales. Marketing sites without direct sales have a harder time with value, but despite its difficulty the issue of measuring the value of a visitor's actions on these sites is not an intractable problem.
Value measurement without direct sales is too involved a topic to cover in full detail here, but at a broad level indirect marketing value can be measured as follows:
- A simple binary score—e.g., as a 1 or 0 for whether a visitor did some desired action or not, such as click through to a piece of content or download a coupon
- An estimated value of some action—e.g., assuming a value of $1.50 for each recipe downloaded from a Web site if research has indicated that each download results, on average, in $1.50 of additive sales
- As a relative value for some action—e.g., if the marketer understands that site registrations are the most desirous actions that consumers do on a site, a registration action can be valued, arbitrarily, at twice the value of other actions on the site
Agility By Example
The reader is probably wondering how the process of dynamic optimization would work in practice. Below, I will outline an application of one potential method for agile Web site messaging optimization, using what I call “the champion-challenger algorithm.” I have selected this algorithm due to its ease of implementation as well as its highly gainful and lightweight nature. As such, it is a strong match for the needs of optimizing messaging for short-term visitor relationships on a Web site.
The champion-challenger algorithm (or c-c algorithm, for short) is an extension of the direct marketing methodology bearing the same name. The champion-challenger methodology works by testing a series of messaging candidates and finding the best performers, which become the champions for a new round of testing where a series of previously untested challenger messages are matched against these champions.
The idea is to keep repeating the process so that the champions get continually stronger with each round of testing and the addition of hopefully better challengers. Implicit in this approach is the understanding that the marketer is using the iterative process as a learning cycle, which will lead to improved candidates over time as more refined messaging insights are gained.
I originally developed the c-c algorithm to facilitate this simple direct marketing process in a dynamic manner on a Web site to optimize page callouts. The ability to selectively display the most gainful messaging as a test is being run was added to make the process more dynamically gainful and autonomous. This added capability allows for the immediate application of gainful insights, as they are obtained, while still facilitating the periodic removal of non-champions and the addition of new messaging candidates.
Next week: How exactly does the c-c algorithm work?
