Mediaplex's MOJO Blog

As a statistician who has spent most of his professional life building & working with predictive models in various marketing settings, I couldn’t help but smile as I read this article on predictive modeling in digital marketing (Diener & Kidd on marketingprofs.com). After all, who doesn’t like hearing their life’s work trumpeted as the next big thing in digital marketing?

Cutting through the hyperbole, it’s true that the rise of “big data” is going to quickly raise the need for “big analysis” … and that’s a lot harder than just shoving data down some magical software solution. I’ve been telling anyone who’d listen for several years now that the growth of technology in our industry to track and store data has far outpaced the growth in our ability to appropriately analyze and understand that data. Many advertisers & agencies today are outsourcing critical quantitative analysis, or have turned to SaaS/3^rd party software solutions to do their thinking for them. That scares the living daylights out of the statistician in me.

Here’s the straight, scary fact: the number of advertisers & agencies out there with the human knowledge base to truly understand the necessary statistical concepts & how to apply them appropriately to the data they are generating is shockingly small. However, the amount of tools that make it easy to “make models” and “conduct analyses” without having any substantive critical knowledge about what these statistical methods actually do is very large (and growing). It’s possible that the growth of the latter is both in response to and the continuing cause of the former. In any case, it’s not a trend I’d like to encourage.

My analogy for you is this – walk into any home improvement/DIY store, and you can buy any number of fancy power tools that make driving nails, cutting boards, etc. easy enough that almost anyone could do it. Does that mean you’d trust just anyone then to build your house for you, simply because they have the tools to do so? I should think not! Data analysis is a contact sport – you need to understand your data, understand your methods, and think critically about how to structure your analyses to get appropriate answers to your business questions. Even if you choose to outsource the dirty work here, someone within your organization needs to be able to speak the statistical language and understand the plan of (analytical) attack.

Amazon’s recent détente on retailers brought up an interesting aside for me. As you may have heard, their Price Check mobile app has caused quite the stir with in-store, brick and mortar retailers. The concept is that you walk into a store, scan the bar code on your phone, and the app lets you check pricing on Amazon and even place an order, and you leave empty-handed out of that expensively-leased building with knowledgeable clerks. The value is that you can hold a product in your hand before you order it, then get best price – something you likely would have done anyway. But in this case, there’s an Amazon kiosk in the palm of your hand.

While it reflects the cruel Machiavellian nature of the web, it’s also incredibly effective, transaction-oriented marketing.

The online and offline worlds have always had a hate-hate relationship. You think back to the days when e-commerce started and we were sure entire stores with shut down. Well, they did. Good-bye, Borders. See ya later, Tower Records. Thanks for all the memories of lost Saturday afternoons flipping through boxes to find the Clash’s London Calling on vinyl. Some of us may remember that before they were called “conversions,” they were called “sales.”

Okay, enough nostalgia… This is a blog about online advertising.

Read more…

Our very own Mark Aronoff was featured on iMedia Connection today in an interview about the lessons and future of online advertising.  Click on the link to get the interview.

A few weeks back, I decided I needed to contribute more around ValueClick Inc.  So, I decided to start a new tradition: Tie Tuesday.  I knew this was a bold undertaking – I’m relatively new, everyone’s at least a few years older than me, and the office dress norm is closer to casual than “business casual.”  Whatever, I went for it.  Every Tuesday, I wear a tie. It’s totally awesome. You should see the amount of respect I get on the Muni.  In the office, my esteemed female colleagues can’t help but swoon as they cruise past my desk to the cafeteria.  I look forward to Tie Tuesday almost more than the weekend. Almost.

Now… around the office, the adoption rate hasn’t been as great as you might imagine.  Some people share the glory that comes along with a sweet piece of neckwear, but I’d estimate that 75% of the team is missing out BIG TIME.

Read more…

According to MediaPost’s Tuesday MediaDailyNews, Time Inc. and Innerscope Research are conducting a study that (at least initially) has found evidence suggesting that “The Millennial Generation” – my peeps – have brains wired differently than our “Baby Boomer” predecessors.

This doesn’t mean we’re smarter, unfortunately. It also doesn’t explain why we complain more than Baby Boomers did “back in their day…” Basically, they’re theorizing that my generation – being brought up in the modern media environment – is wired to consume media differently than those brought up in the archaic print/broadcast environment of yesteryear.

Read more…

Within the digital ad space right now, the hot topic is around attribution analysis. For publishers, it’s hot because they want credit for key engagements with a prospect. It’s hot for advertisers because they want to better understand how to maximize their ad spend and their engagement cycle.

But it wasn’t always this way. Not too long ago, we spent very few cycles on how people react to a marketing mix. We thought it was about awesomely influential ads on high-end sites, which required you to put all your budgets on introducers (first click) and closers (last click). Now we know that only focusing on those are not only unsustainable, we know there are many more ways to optimize.

To be timely: attribution analysis is like Moneyball, but for advertising.

Read more…