Grace Huang is an ex-McKinsey consultant whose new startup, DeepZero | iPinYou, has set its sights on the Holy Grail of behavioral targeting - something that, as the interview makes clear, has not exactly taken off in China. I asked her why that is, and about other issues related to the Internet ad market in China. Here are my questions and her emailed responses. For you Ogilvy Beijing employees, Grace will be the speaker at our Innovation Day brown bag lunch series this coming Wednesday.
Kaiser: In the U.S. and other developed Internet markets, behavioral targeting has been a really hot topic in the last 18 months, with companies like Tacoda and Blue Lithium being acquired by big players like AOL and Yahoo, respectively. What’s keeping it from taking off in China?
Grace Huang: In my opinion, it’s hindered both on the supply and demand sides. On the advertisers’ side, Internet marketing is still new to lots of bigger companies and agencies. Most of the buying and planning are limited to safe and traditional portal/vertical sites. Compared to 17% in the U.S. markets, spending on internet marketing in China represents only less than 4 cents of the total advertising dollar. Moreover, online advertisers are just getting used to online advertising with display advertising and search marketing – and even display advertising is still in the early stages, given that much of it is sold on a CPT [note: Cost per Time] basis. On the other hand, behavioral targeting companies in China are just starting up. I don’t think many people have figured out the essence of behavior targeting and especially how it can be applied to the current China market. In the U.S., many companies have thought hard about how to make behavioral targeting a real solution for advertisers. It needs a combination of strong business operations and technology advancement. So it’s not an easy thing at all. Even in the States, it is only recently starting to get real momentum in terms of proving it yields better ROI, creating robust profiles and algorithms, etc. So it’s a process of evolution, both for the advertisers and the suppliers. Just as it took a while for the behavioral advertising market to develop in the US, it will be here as well. However, it will go faster in China because 1) we have the demonstrated success from the US, and 2) the Chinese have proven to be quicker to adopt new advertising techniques, as was evidenced by search marketing.
The good news is that we do see the needs on the advertisers’ side for a more efficient way of spending money online. Based on our talks with many advertisers, we found that brands have realized that compared to the size of the Internet population and the amount of time people spend on the Internet, the advertising dollar is under-spent. The dilemma is that they don’t know any solution that can optimize their display ads and generate higher efficiency. However, advertising on high-traffic portals for many targeted brands is not as efficient as desired. Search marketing is great, but it only reaches people who have typed in the magic words. It captures consumers who are already in the final purchase stage.
Kaiser: Is contextual advertising more appropriate for China at this stage? Why haven’t we seen particularly successful examples of contextual ad serving in China either?
GH: Contextual advertising works well when the content is commercially related. The problem, however, is that only about 20% of Web content has commercially relevant content, according to research by Tacoda’s investor Masthead. Associating a text link with a site that reviews consumer electronics is likely to generate a “clickthrough” to an electronics retailer, but consumers spend much of their time online looking at news, sports, and weather, none of which have many keywords that are going to be of much value to that same electronics retailer. Attempts to use contextual targeting with news have had the unfortunate result of promotions for Samsonite being placed next to stories about suitcase bombs and offers for life jackets being placed adjacent to stories about capsized ferries.
Contextual targeting works well in certain types of consumer content. A consumer researching the purchase of a new car, or computer is likely to find a lot of relevant advertising. But contextual networks can only target that advertising when the consumer is on the page with that content. If, ten minutes later, he or she is browsing the local sports scores, he or she will see nothing related to an earlier search. Therefore, in our definition, it is still “reactive” advertising.
Kaiser: We’ve all heard it said that Chinese Internet users aren’t as obsessed with privacy as their American counterparts. Do you find this to be the case? And what does that mean for behavioral targeting in China?
GH: Well, overall speaking, I would say yes. However, this does not mean companies like ours should take advantage of this. We are fully aware that in the U.S., privacy protection has become the number 1 concern for BT players. In China, we want to become the champion of discipline in protecting privacy. Therefore, we will always avoid collecting any personally identifiable information about consumers. We are also thinking of various ways for consumers to opt-in or opt-out. We think privacy protection should always be the top priority for companies in the advertising area.
The other point I want to make is that we believe behavioral targeting is intended to offer consumers more relevant advertising (since advertising in some form is unavoidable anyway). Therefore, as long as privacy is well protected, consumers will welcome such a form of advertising.
Kaiser: When I talked to you for the first time, we talked about the trade-off between reach and precision: The more precise the targeting, the narrower the reach, typically. How does DeepZero | iPinYou plan to get around this problem?
GH: The idea is really to understand consumers based on behaviors and profile them based on accumulated inputs. Instead of being reactive, we need to be predictive. Instead of indexing pages, or keywords, we are indexing consumers Instead of creating preset segments, we understand multiple dimensions of consumers. As a result, we can deliver ads to the right consumers regardless of where they are. We target a consumer not because he is now visiting a page, but rather because he has searched some keywords, and he has visited certain sites, etc. As a result, different people see different ads when they log on to the same page.
This creates additional benefits: we can reach targeted consumers on more inventories and therefore expand the reach of other targeting mechanisms. For example, our Vertical Cruiser solution helps brands not only reach people on the relevant vertical site when the brand advertises on that site but also reach people after they leave that vertical site.
Kaiser: What’s DeepZero | iPinYou’s long-term vision for what it wants to enable advertisers and publishers to do?
GH: The long-term vision of DeepZero | iPinYou is to allow advertisers to be able to come in and define the specific target consumers based on different dimensions for an advertising campaign. DeepZero | iPinYou’s system will profile consumers based on multiple inputs, including demographics, psychographics, category interests, lifestyle, and purchasing funnel. Through DeepZero | iPinYou’s network of publishers, the ad will automatically appear whenever the target consumers visit the page. What ValueClick recently launched in the U.S. “precision profiling” is similar to what we have in mind.
The benefits? For advertisers, we see a higher conversion rate or more targeted exposure to advertising. In one of the campaigns we run for an English education brand, the click-to-conversion rate is almost 3 times higher than the result generated through traditional single-site advertising. For brand campaigns, we help brands to reduce the effective CPM.
For publishers, we will see much higher revenue per visit. Also, much-untargeted inventory will be turned into targeted inventory, therefore getting a higher unit price for publishers. There is an additional benefit for consumers as they will not be bothered by irrelevant advertising. We want consumers to see the ads they want to see.
