Peggy Anne Salz at MocoNews reports on Tom Wheeler’s recent article in TMCnet “Mobile Advertising: The Hidden Breakthrough“. Tom talks about how datacasting can be used to lower the cost of mobile advertising. Rather than deliver advertising to mobile 1:1, deliver advertising one-to-many. Peggy’s article, and the discussion that follows, argues about whether the commercial-like ad model will prevail, given that people are tuning out mass market advertising with DVRs today, and than mobile demands more personalized advertising.
I think this is a common misconception based on not understanding how datacasting can work. With datacasting, you can set up a data carousel that broadcasts dozens of advertisements targeting any number of demographics or market segments. When the client (the phone) sees a package that it deems relevant it actually stores that package on the phone.
This works well if the customer base can be segmented into a reasonable number of groups. I think this is the real point of contention. Because mobile phones and servers have the potential for what I call hyper-personalization, people believe that the market (the advertisers) will take advantage of that capability. To this, I am quite skeptical.
I think there are going to be five types of advertising to hit mobile phones:
- mass-market advertising
- subscription advertising
- personalize-later advertising
- niche advertising
- hyper-personalized advertising
Mass-market advertising is essentially what we all know and (cough) love on television today. Typically, its brand advertising that is trying to reach the mass market with little regard for demographics. Seth Godin talks about this type of marketing in an excellent blog post entitled “Reaching the Unreachable.” We can argue all day long whether or not its appropriate for mobile, much less today’s consumer, but you will find plenty of advertising in this category — especially in the early days when we see “sponsored” channels of content. Datacasting is very cost effective for this type of advertising.
Subscription advertising is when the consumer tells the service or advertiser that they are want to get certain types of advertising. Typically, the consumer will indicate that they are interested in a specific category (automobile advertisements, for example), but it can extend down to the brand level (BMW advertisements, for example). For obvious reasons, self-selection is very valuable to an advertiser. At the category level, datacasting is very cost effective for this type of advertising.
Personalize-later advertising (not the best choice of terms, I admit) is when the advertiser sends down a “generic” advertisement but that advertisement is personalized after it hits the handset. An example of this would be a generic BMW video advertisement that gets sent to the phone and then the client (either in response to a user interaction or not) fetches directions to the nearby BMW dealers based on the consumer’s location. Datacasting is very cost effective for the generic form of the advertisement, using HTTP for the personalization.
Niche advertising is when advertisers target a highly specific demographic to the point that it appears to be 1:1 targeting, but in reality is not. Niche advertising is fairly common in the cable television market where there are channels very specific to a customer demographic (a jewelry channel or men’s outdoor life). Datacasting is very cost effective for this type of advertising if the advertising can be multiplexed with the channel’s normal video programming (i.e., its leveraging the channel’s existing bandwidth).
Hyper-personalized advertising is the truly 1:1 targeting of an advertisement to a subscriber. Typically, a profile is built describing that subscriber and advertisers can then target profiles that meet their needs. Datacasting is not very cost effective if it is truly 1:1, but my expectation is that most hyper-personalized solutions will actually behave more like “niche advertising” in order to get scale.
I am in complete agreement with Tom Wheeler when it comes to the effectiveness of datacasting for mobile advertising. As I hope that I have shown above, there is not just one paradigm for mobile advertising. We might not know the rate of return for each of the methods I outline above (as measured by click-through or purchases), but given that four of the five are enabled with datacasting, I think thats a good approach to take.
June 27, 2007 at 8:34 pm
Ted,
Do not quite understand why datacasting or filecasting is needed for mass market and other forms of channel based advertising when these ads can be be sent down in a cost-effective way from the network.
Datacasting does not solve the whole nine yards in targeted advertising alone. Pre-caching of ads and play back is not a new concept (being done in TV today, the likes of Visitble World and Invidi), but getting the business model right is the key to a sustainable and progressive Mobile TV ecosystem.
Srini
June 28, 2007 at 7:27 am
As you point out, getting the business model right is key. Cost is a key element of the business model — you normally don’t want costs increasing at the same rate you increase consumption. As you also point out, cost effective mechanisms are needed to delivery advertising. My belief is that datacasting is one means of achieving cost effective advertising.
I looked at Visible World and it appears to represent the “personalize later” mechanism for advertising that I describe above. “Trafficked as a single ad, these dynamic elements are assembled automatically at airtime to create a spot that’s exactly right for its time, place, and audience” To me, this is one form of datacasting.