Building from the Ground Up: Lessons from My Monstrous CD Collection on Data Segmentation and Multi-Channel Optimization
If you were born in the 1980s (like me) you came of age at a time of disruption for the music industry. In the 1990s record companies were slow to adapt to new options for streaming music online. My earliest memories of this were with the music-sharing sites Napster and Kazaa.
These pioneers may have had their legal issues, but the industry was disrupted and there was no turning back: the days of record companies selling entire albums of songs to consumers were numbered. MP3s began to replace CDs (remember CD stores!? My local store was called Spinners…). The idea of purchasing 10-15 songs at a time when you really only wanted one or two began to seem silly. Soon iTunes picked up the torch from the Napsters of the world and purchasing individual tracks instead of albums became mainstream.
Fast forward to my high school days: among my friends I was the slowest adopter of this new way of enjoying music. I had a certain (looking back, embarrassing) amount of pride in the number of CDs I owned. And it wasn’t just that I preferred CDs over MP3s —it was worse. I was very impressed with myself for purchasing a band’s entire album. I was enjoying the music the way the artist intended—and making sure they were paid fairly for that enjoyment. Downloading individual MP3 files of music was simply not for me. My goodness I must have been insufferable.
I’ve been accused once or twice of being a slow learner. As my career in multi-channel analytics has progressed I am proud to say I’ve finally learned the lesson that came so naturally to my high school friends: when distributing a scarce resource, artificial rules or “quotas” are anathema to optimization.
Record companies were slow to learn this lesson: they forced consumers to buy 15 songs when they only wanted two, giving Apple room to disrupt them with iTunes. The artificial construct (the album) forced a sub-optimal allocation of resource (the $16.99 for a CD) and ultimately hurt the goal of the enterprise (consumer enjoyment).
We often encounter multi-channel retailers making similar mistakes when spending advertising budgets. Here are the three most common occurrences with advice on how to avoid being a 1990s record company:
Applying top-down quotas when purchasing prospect records from data providers
- Description: Many circulation directors pre-determine the % of a prospect campaign that will be sourced by Provider X vs. Provider Y. Quota systems lead to making selections regardless of the data source or segment’s inherent value—quantity is needed to fulfill the obligation. This almost always causes clients to select inferior records from one source when the other source had better options.
- Antidote: Build prospect campaigns from the Bottom Up (BUMP, or Bottoms-Up Mail Plan). Individual data sources or segments should “earn” their way into a campaign based solely on their performance with no consideration of their supplier. Only once quantities have been reached using this bottoms-up methodology should clients check the ratio of the campaign being supplied by Vendor X vs. Vendor Y. In this case the quota-by-supplier is the artificial construct preventing optimization.
Building out Pay Per Click Ad Groups with more than one keyword.
- Description: Nearly all of our clients bid on 10-30 keywords per ad group. This blends their CPCs and the performance of both the weak and strong keywords into a deceptive average, causing clients to miss out on under-funded keywords that deserve more budget.
- Antidote: Single Keyword Ad Groups. SKAGs have become standard practice for MeritDIrect’s SEM Management department; they force each keyword to live or die on their own. Keywords s cannot get a free ride by “hiding” in ad groups with stronger performers. Strong performers, for their part, are recognized faster, shrinking optimization timelines. In this case the multi-keyword Ad Group serves as an impediment to optimization—get rid of it!
Setting dial quotas across telemarketing campaigns
- Description: A common way to manage telesales reps involves enforcing dial-quotas by campaign. This is used to enforce discipline in what can be sometimes be an unruly medium. For example, in a week a rep may be required to make 100 dials in a Reactivation Campaign as well as another 50 in a Cross-Sell Campaign.
- Antidote: In this case our artificial construct is the campaign—strip it away and we have a requirement that a rep make 150 total dials. The universe of Reactivation candidates and Cross-Sell candidates should be combined and modeled for responsiveness. One scoring system then dictates the rank-order of dials for the rep and may result in a distribution that is 120 dials to Reactivation and only 30 to Cross-Sell. Eliminating the campaign leads to optimization.
Best-in-class Multi-Channel Retailers understand that artificial constraints or quotas lead to spending money unwisely. Don’t be stuck with CDs in the 1990s when setting those Ad Budgets!
Contact me today for more information on optimizing your Multi-Channel Marketing efforts, or if you want any of my old CD’s.
Vice President, Business Intelligence + Analytics
Interested in more content from our BI + Analytics Group? Check out more content below!
Maria Jaime on Visibility Beyond the Conversion
Andy Joyce on PPC
Andy Joyce on Cross-Channel Optimization
Maria Jaime on PPC