Jonathan has helped optimize...

« Jay-Z & Rocawear Lay Down the Viral Video Marketing Beat | Main | Ad Testing »

A Digital Marketing Lesson from Wall Street’s Woes

DiceAs the New York Times points out in Sunday’s Business section the recent collapse of the stock and bond markets has much to do with the automated black-box systems of hedge funds whose algorithms are seeing behavior that despite all their statistical modeling has no precedent. The automated systems have failed and it’s now people that need to come to the rescue.

From the article The Week When Risk Came Home to Roost by Gretchen Morgenson:

None of this would be a problem of course if fund managers were not relying so heavily on precedent to make their decisions. Computer models seem so perfect, so scientific, so flawless and they are advertised in precisely that fashion. Ingenious models lull investors into a dangerous complacency…

Just like in the investment markets black box algos in the marketing world will lull marketers into complacency and will never improve upon human results. This should be a reality check in our industry around the current buzz and euphoria for the automated behavioral targeting and "set-it and forget it" suggestion tools getting loads of hype and capital.

Why aren't these automated systems always going to work?

Having done hundreds of tests I can tell you from personal experience that just like the investment markets so much of what is seen in consumer behavior is counterintuitive and could not have been predicted.

Financial markets do not always trade in a way that is typical or predictable. And when they deviate from the norm, all the wonderful and smart trades stop behaving according to plan.

C-level execs and other decision makers investing in these platforms should understand there will be risks risks associated with them. The risk will be magnified if they are depending on these automated technologies to replace other tools and technologies controlled by the human marketer.

Every marketer would love to just sit back and collect great results - but just like Wall Street - in the markets where we play risk is never gone because while behavior is measurable and quantifiable, it is never predictable.

That's why they call it marketing.

TrackBack

TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d8341ccaa353ef00e39826ee198833

Listed below are links to weblogs that reference A Digital Marketing Lesson from Wall Street’s Woes:

Comments

Jonathan,

This is a great post highlighting two valuable lessons that we do not seem willing to learn: that technology is a means and not an end, and, that we are unable to accurately predict behavior – and hopefully never will be.

I think one of the paragraphs in the article you reference exposes a basic flaw in the plan. Particularly, “More simply put, the goal is to transform advertising from mass messages and 30-second commercials that people chat about around the water cooler into personalized messages for each potential customer.”

Isn’t it desirable for people to chat about our message around the water cooler? A goal of every promotional campaign should be to create some sort of buzz. With mass advertising, the conversation starts because everyone gets the same message – we can all join in because we know exactly what the conversation is about.

Will the buzz start and propagate as easily if we each receive a unique message?

Cheers,

Stephen Da Cambra

Jonathan,

Good post, and I read it -- and immediately thought about a myriad of responses to write...but while your premise is correct -- I do believe that a machine, and a proper working "black "box" as you describe it, (if built properly and yielding correct data) can far outweigh, anything a human being can decipher.

From a creative standpoint, you are 100% correct.

From a "arb" or risk opportunity, in looking at search/display media - there is such a huge amount of data that flows through a system, and needs to be decisioned on, in REAL TIME - it is staggering.

Lets look at a few elements of banner/display data points to be decisioned on.

1) most optimal ad (if multiple per creative)
2) creative
3) size
4) Landing page (if being optimzed)
5) user preferences and previous history
6) frequency cap
7) potential decision to DECLINE the impression opportunity

This all happens in less than .05 of a second -- and outside of directly hardcoding creatives into a particulair page (like Yahoo's front page for a day) - there needs to be some kind of "black box" intelligence around a few critical points.

1) yield management
2) segmentation
3) decisioning

Your example cited above about Wall Street, is excellent but flawed - (and I read it as well but cringed when I read it).....

With huge amounts of data, isnt a correction or a "flawed data set" bound to happen - outside of the bounds of normality?

I cannot make excuses for people trading in excess, nor can I make excuses for Hedge funds who relied on GREED and profit -- more than protecting capital....BUT....

Isnt today's Hedge Fund, kind of similar to MySpace and Facebook, (as it relates to data and page consumption) versus portals and search data.

We should never "set it and forget it" but at the same time, linear programming built to "spit out" important data points on driving results for both publishers and advertisers, are the building block for moving the market forward.....

Corrections in markets, and flawed data happen in all marketplaces (including equities as well as internet marketing) but the key is make sure your "black box" (as you describe it) is scalable and built to WITHSTAND flucuations......AND CAPTILIZE ON DOWNTURNS, and mitigate risk in upward swings.

The contrarian I am looks at downturns as a "buying opportunity" rather than a reason to fear.......

Much like so called "remnant inventory" in online media...I look at it as a buying opportunity - to yield our intelligent decisioning to "arb" a HUGE opportunity.

It's part human, and part machine.

But for many it is 60% machine (real time data) and 40% human (emotions)....

Disclosure: (short: Humans)
(long: intelligent machines)
(long: Livingston)

Rock on,

Andy


Stephen- You raise interesting questions that I can't answer but at the end of the day buzz is only helpful if that added awarenes generates higher ROI. These days buzz is viral and maybe the buzz meter will be higher if the messages are more relevant?

Andy- hard to argue with the most intelligent post here in sometime. so yes my arguement is flawed but if it was able to elicit that response than at least it more than served its purpose. :)

Oh, some dizzy!

Hi, some data on the 05wl.nets, come from China!

The comments to this entry are closed.