Magnetic CEO Josh Shatkin-Margolis spoke to AdExchanger.com recently about his company and the search retargeting data space.
Click below or scroll down to read more:
- The Company's Evolution
- Data Marketplace Observations
- Target Market
- How Does Magnetic Handle Creative
- Future Product Development
- Comparing Financial Markets To Ad Marketplaces
- The Privacy Debate
AdExchanger.com: Since we spoke a year ago, how has Magnetic evolved?
JSM: What's evolved over time for the company, is that we were initially selling data similar to some of the other data exchanges. Specifically, we were selling to ad networks, DSP's, and other ad exchange buyers the ability to target a user based off of their search history.
What we found over time is that while we've seen good results selling data, doing media optimizations is difficult, overlaying data is even more difficult, and overlaying an extremely granular form of data like search data is even more difficult.
So, we started selling data plus media directly to agencies and advertisers, and that has been the next step within the company and we've seen much better results consequently.
And the reason why is that our delivery managers, our campaign managers, our traffickers, have built up expertise in running search targeted display campaigns.
I think the data market is like any new, growing market and has its challenges. It's obviously still an exciting premise to be able to target users and display media with a tremendous level of transparency and efficiency. I think some of the challenges that the data market has seen is that search re‑targeting is dependent on how DSP's, ad networks and the trading desks optimize data.
Another challenge is figuring out the appropriate pricing model. Buyers are moving from pixel stamps to media‑based pricing and making sure that they're only paying for data as they're using it.
Figuring out what's an appropriate price of data relative to media is a challenge, too. Initially, people were very excited about it, and prices were a bit higher. Since that time, prices have pulled back a bit.
That has not changed at all. We’ve tested other sources of data. There has been some interest in collecting data from referring URLs. But, Magnetic does not collect data from referring URLs. So, some of our competitors go after data that can be collected by their widgets which show up on web pages, and those widgets are able to detect the referring URL and the search.
Magnetic does not collect data, because after we tested the data, we found that the data was of dubious value, and the reason was that the data selects for non‑intenders.
For us, we found that consistently getting the data from second tier search engines, e‑commerce sites, comparison shopping is best and there's no skew in the data in any direction.
We continue to sell data to all the ad networks and DSP's and ad exchange buyers. We've seen tremendous growth working directly with the agencies and advertisers on the display side and have built up that expertise. We've seen great progress, the best results in education, credit related products, auto, and telecom.
Similarly, having built up this expertise from a performance standpoint, we've since launched an SEM‑like interface where a search engine marketer can come in and create campaigns, ad groups and text ads and pay us on a price-per-click basis in a familiar way to working with a search engine. They can buy SEM‑like inventory through us.
Instead of paying $25 a click on car insurance to have their ad show up alongside a Google search result, they can pay us one or two dollars a click to have it up show up targeted at the same audience but after they've left the search engine.
On the creative side, our whole business is enabling as much search re‑targeting as possible, and the way we're doing that is giving very similar interfaces to the major groups that want search re‑targeting. The three major groups are the people that sell media, which would be ad networks, publishers, ad exchange buyers and DSPs.. The second group are people who are display advertisers and are used to high-touch models. They sign insertion orders and give us creative and have a campaign management team actually deliver those campaigns for them.
And then the third group is search engine marketers who want to extend their search marketing campaigns and drive traffic at a lower CPA.
Now we do serve text ads for those search engine marketers because it's easy for them. We're not really banking our business on the merging of the search industry and the display industry in that these two industries are still fairly siloed.
One of the mistakes that I made when I first started the company was that I thought since search is a 20 billion dollar industry and display is a 10 billion dollar industry, you could put them together and have a 30 billion dollar industry. The reality is if you put them together you have an intersection, which is a 300 million dollar industry.
I've always thought the holy grail is you find a search engine marketer who manages the keyword list and a display advertiser who manages the creative with a perfect call to action, the perfect enticing graphic - and you combine all those things together. I still believe that it's the holy grail, but right now most of our display advertisers do not have a keyword list, and most of our search engine marketers do not have graphical ads created. But, the performance they're seeing even without those additional interactions is still fantastic.
Yes. We absolutely can evolve the product and we're absolutely integrating with mobile and video. We integrated with Bright Roll last year and we're integrating with mobile companies this year.
But, we're not integrating with video search as a data source. I think we're very cautious in terms of where we get data from and if someone’s on a site searching for a video it doesn't necessarily demonstrate a consumer intent.
I think it's valid in a higher level sense in that there's a lot of data you can analyze and provide some value in that sense - similar to perhaps a mutual fund sitting on top of a stock market similar to how a demand-side platform (DSP) is deciding which media's good and which media's bad. I'd say that's where, in my opinion, the comparison ends.
Stocks and media are very different. When you buy a share of stock, you're not really doing anything with it. When you bought a share of Microsoft stock that actually meant that you were some how physically associated with Microsoft. Meaning that your brand was associated with them and people knew you were associated with them and it affected your business and it affected your ability to execute on your business. Media is very different. When you buy and sell media, these advertisers are putting their brands on these websites, it's not as simple as just trading a commodity.
For Magnetic, it has not impacted us and the reason is we removed ourselves from the privacy debate by building a very different system. We're laser-focused on keyword targeting. We're not doing user targeting, so we're not building profiles. So, we're aware of the privacy debate going on, but it's really more targeted at the different behavioral targeting systems that are building profiles on users.