Tech Mentat

The aptly named ‘Does Demographic Targeting Matter?’ is a session devised to tackle the controversial question of the efficacy of Demographic Targeting (and I would add… of targeting in general). On a very high level we ask, ‘Is demographic targeting a gimmick we use to legitimize intelligent web marketing or goes it have potency in its practice.’ Personally, while I am a rainy-day-weather anal retentive type, I tend to believe that numbers and segmentation (demographics) drive success – yet numbers limit us because they lack a global understanding (i.e. numbers can only speak to what they know/report upon but if you target, you obscure/limit numbers and miss a larger picture).

But let’s see what the pros have to say…

First up is Jed Nahum who is the Director of Product Management for MSN Search. He came today to talk about the MSN AdCenter and the power of using the AdCenter’s demographic information. I was skeptical at first but I have to say, the presentation wasn’t a total sales pitch of MSN AdCenter.

After running through a brief overview of MSN AdCenter and it’s current status (MSN was down all weekend for a massive update), Jed went into the subject of How do they target demographics? Unsurprisingly but equally frightening, Microsoft gathers user information through windows registration, windows live and passport (think hotmail) registration. Through this information they know pretty much everything about you that is worth knowing for a marketer like myself (in terms of demographics). The data is a lot more trustworthy than basing it on IP address and cookie information – it’s a deeper, more powerful conversation MSN is having with its users (*without them knowing about it). Search data per keyword is kept in through a cookie – Jed made a point of stating no delicate information is stored in this cookie, just age and group information. As a final point, Jed pointed out data that showed their search demographic information (the information they report on per keyword) is consistent with their passport information. Meaning, the information they collect is truthful and thus useful for marketers.

The final part of Jed’s presentation focused on the usefulness of the demographic information – moreover, how MSN AdCenter can benefit you and why you should be using the AdCenter. I have summarized his basic points below:

• Demographic information used for keyword sets and desired audience. Age and gender distribution for a given term.
• Demographic information allows you to target your ads/keywords better. Quickly speaks about bid boost – boosts bid when your demographic is searching for your terms.
• Gives a few examples of demographic information you could use. Purses – mostly women, Tag (body spray) – mostly men, Kitchen appliance – 50/50 split, Viking kitchen appliance – mostly men. <- This suggests a way to alter your keyword buy OR your ad copy.
• States that because of market intelligent, marketers are seeing a higher ROAS, better conversions.

Kevin Lee, the Executive Chairman and Co-Founder of Did-it.com, hits the floor next.

Introduce a few case studies on demographic targeting.

Kevin starts by asking the question Why Demographic targeting? He states that it is a better way to get the right message in front of the right user. This simple point is elaborated upon to present an ‘everyone wins’ situation – the advertiser sees better ROI/ROAS, searchers are happier because they are given a product that more directly related to what they are searching for and publisher more money the CPC’s and CTR’s are higher. Pleasant but I wasn’t buying the farm…

Kevin then continues to speak about knowing your demographic information. He coins the term Power Segments – a phrase that is meant to describe your companies demographic where you yield the most return (whether that goal is sales, traffic or page-views). He suggests that many web marketers who start to use MSN AdCenter (which is currently the only online ad platform that allows for demographic targeting) have not yet defined their Power Segment. Thus using MSN AdCenter’s demographic targeting will inherently be a fruitful endeavor. Kevin states that you need to know your users - through registration, form submission and collecting user information. Once you know this information and you know which of these users you are seeing the highest return from, you should focus your efforts on your best customers (not all your customers) and know your performance values – conversion, immediate value or long term value. Seems like a simple enough point but a great one to be made AND one that I (personally) think run adjacent to the idea of the ever popular long tail strategy (where there may not be a power segment or core).

Kevin quickly touches on using geographic information wisely –bid boosting on different areas of the country and having a different message and bid price for areas of the country.

Kevin’s final point is to state that you need to be cherry-picking your clicks. If you use demographic information, find your power segment and advertise accordingly then before you purchase your clicks you will know what the value of that click is and you will only be purchasing the clicks that are converting most. Kevin’s message is to use your demographic levers to hit your power segments – peel back the layers of the onion. This works best on cusp keywords (keywords that perform best in position 3-7) – you could kill your bid for most points of the day and then jump to position 1 at a given time of the day.

The final speaker was Michael Sack, the Executive VP & Chief Technology Officer of Inceptor. Michael is a lively guy and a refreshing voice in the sea of otherwise monotone voices/presentations. Michael, admittedly, is the naysayer of the group and so he did a study to see if demographics really matter. His initial thought is that it [demographics] might be a gimmick and that it may not work for web marketing the way it does work for offline marketing (though it should be stated that offline/print marketing demographics are also gathered and summarized FAR differently).

It Depends! (this was his triumphant first remark)

What Michael explains is that it depends on what are your goals – brand, drive traffic, do you care who is buying, are you on a limited budget or are you only paying-per-action [performance]? However, simple this may seem, Michael brings up a point that others seem to neglect and that is – you are selling a one-size-fits-all ad platform (chanting, you must use demographics) while preaching that advertising and sales should not be marketed as if we live in a one-size-fits-all world.

Michael then gives us a quick history lesson, telling us that demographic targeting started as a way to better target your potential advertisers (TV stations went out and approached potential advertisers and used demographic information as a way to sell to them… not the other way around). He also states that in offline advertising, when companies target demographically the advertiser is getting a lot of overlap. You may target an Ad during Seinfeld that is meant to hit a demo of 25-35 year old, upwardly mobile individuals BUT you are still going to get the 65 years. He states that is why people say ‘half of all advertising works – we just aren’t sure which half works.’ Finally he makes the point that offline is a PUSH world (you are pushing your message) and online is a PULL world (you are sucking users in).

Kevin lightly hits upon the idea of privacy (which I didn’t think was too relevant) and asks the crowd: Do you like the fact that Microsoft has your info? Most people don’t (in the crowd and in general), thus he states that people set up false profiles. Does this help your targeting?

Finally, Kevin gets to his test. His test consists of three parts:

1. He ran general ads
2. He ran ads with demographic targeting
3. He ran ads and manually optimized

Kevin’s results (from a test that I don’t think ran long enough) were that when he applied demographic targeting to the ads he was running on ‘General’ he saw a drop in cost, his average CPC boosted about $.10 and his conversion rates spiked. When he applied manual optimization to the ads he was running on ‘General’ he saw a drop in cost (not as much as with Demographic), his conversion rates spiked (not as much as with Demographic) and his revenue was higher than it was with Demographic targeting.

His final assessment was that demographic targeting does work but you are sacrificing traffic. Employing demographics depends – if your goal is acquire customers then NO but if you have a niche market and a limited budget then YES.

I like Michael’s enthusiasm and I like what he was trying to accomplish, however, I think Michael left out some obvious details and this study felt like it was put together last minute (the good old college try). I learned from his questioning but not from his results…