Use Dynamic Message Optimization To Increase Campaign Response Rates

April 15, 2009
138 Views

Imagine the following scenario:

  1. 8:00am: You kick off the day with an online ad campaign using a mix of creative treatments.
  2. 8:30am: You detect that one of your creative treatments is generating 3x the response of the other treatments.
  3. 4:59pm: You realize that you should have turned off all of the failing creative treatments – leaving only the top performer – hours ago!
  4. 5:00pm: Your campaign completes having delivered merely average performance.  Feeling stupid, you double-check the requirements for the Darwin Awards – just to be sure this doesn’t qualify you – and breathe a sigh of relief since your incompetence has neither left you dead nor rendered you sterile.

How much time should have elapsed before the failing creatives were turned off?  As little as possible! And that is yet another reason why data load and query performance matters.

One solution here involves a campaign manager simply checking a report, identifying the top-performer, and killing all the non-top-performers (note here that even this approach requires timely and accurate data to inform the decision).  But a growing number of firms are moving away from manual tuning by campaign managers

Imagine the following scenario:

  1. 8:00am: You kick off the day with an online ad campaign using a mix of creative treatments.
  2. 8:30am: You detect that one of your creative treatments is generating 3x the response of the other treatments.
  3. 4:59pm: You realize that you should have turned off all of the failing creative treatments – leaving only the top performer – hours ago!
  4. 5:00pm: Your campaign completes having delivered merely average performance.  Feeling stupid, you double-check the requirements for the Darwin Awards – just to be sure this doesn’t qualify you – and breathe a sigh of relief since your incompetence has neither left you dead nor rendered you sterile.

How much time should have elapsed before the failing creatives were turned off?  As little as possible! And that is yet another reason why data load and query performance matters.

One solution here involves a campaign manager simply checking a report, identifying the top-performer, and killing all the non-top-performers (note here that even this approach requires timely and accurate data to inform the decision).  But a growing number of firms are moving away from manual tuning by campaign managers toward automated optimization algorithms in order to reduce decision latency and increase decision quality.  The result is increased campaign response.  The figure below, adapted from material in Tom Davenport’s “Competing on Analytics”, depicts this analytic maturity curve.

Analytic Maturity Curve

In addition to the increasing number of digital media firms building custom solutions to address this optimization problem, there are also a growing number of marketing service providers focused on it – SnapAds, Teracent, Dapper, and Tumri come to mind – as does ValueClick ActiveAds (whose offering is described in this MediaPost article but not yet featured on their company website).  Each of these companies focus on optimizing a campaign’s creative while it’s in flight to better ensure that the right message gets delivered to the right person at the right time.

Optimizing the "right message" based on the target audience's position in the conversion funnel

As the image above depicts, it’s a good idea to optimize messages for a given target audience based on their location in the conversion funnel.  However, recent research from Yahoo! indicates that some classes of consumers (in this case, tech enthusiasts) no longer follow the traditional purchase funnel.  Instead, they “chart personalized purchase trajectories that work right for them”.  You can download a PDF of the “Tech Enthusiasts Study” here.  This research presents just one more reason to dynamically optimize your messaging – just as you dynamically optimize your segment definitions.

While there may be differences in the approaches taken by the marketing service providers listed above, their value propositions are all based on the same principle – time is money.  The figure below is adapted from a presentation by Richard Hackathorn and illustrates this idea:

Time is Money

The relationship between performance – minimizing data load and query response times – and business value is incontrovertible.  Understanding this relationship – and acting on it – just might let you forget about those Darwin Awards.  It might even put you on the path to winning the awards you really want, like the CLIO, Addy, IAC, Webby, The One Show, and OMMA awards…

Photo credit:  TheBuiBrothers.com