If your job involves online advertisement placement, you’re likely intimately familiar with contextual advertising on traditional search engine results. But recent research indicates that new “realtime” search engines could be the wave of the future for smart ad buys. According to these reports, users of realtime search engines search many more times per day, per query than they do on a traditional engine. This means that with a realtime engine there would be many more opportunities to serve ads to the same user.

For both realtime search engine businesses and advertisers to capitalize on these findings, new and improved realtime monetization models are needed. This is the focus of Web 2.0 Expo’s session Monetizing the Refresh – Advertising and the Realtime Web. This session – taking place Thursday, 11am in room 2010 – features Anamitra Banerji (Twitter), Tobias Peggs (OneRiot), Eric Picard (Microsoft), Brian Pokorny (SV Angel LLC), and Christina Wodtke (MySpace).

OneRiot President Tobias Peggs spoke with us about this panel, monetizing on realtime search engines, and what advertisers need to know.

Kaitlin: I had a discussion the other day about what exactly “realtime” means as the term as applied to Web apps and sites since it has evolved over time. What’s the current meaning and do you foresee there being anything more realtime than what we have now?  What would that look like?

Tobias: OneRiot organizes the “realtime web.” We think of the realtime web as the content that is really resonating with people on the social web right now. In realtime, millions of people might be Tweeting links to interesting web pages, or posting videos on MySpace, or “liking” news stories and syndicating them to their Facebook profile… All of this realtime social activity points to content that is resonating with people right now. We index that content and allow you to search it. In other words, we’re organizing the realtime web and – through a search interface – helping you to discover the signal in the stream of social noise, to find out what’s going on right now.

That is a very people-centric view, and distinct from the view that “realtime” needs to mean “recently published.” To us, “recently published” means a firehose of information that is often too much to be consumed by any sane person, and frequently is not relevant.

A good illustration of this point is the event of Michael Jackson’s death. Once TMZ had published the first story, every news organization on the planet soon published its own take on the pop star’s demise. A view of the realtime web that says “recently published” is king would have uncovered thousands of similar news stories telling you that MJ had died – not very useful or informative, especially if you’d already seen the TMZ story. But a people-centric view of the realtime web would uncover the content that that was really resonating with people at that point in time. Of course, that would include the TMZ story, but also YouTube Videos for “Thriller” or “Beat It”… content that might be years old but content that, at that point in time, was really relevant to people as they were thinking back on Michael Jackson’s life.

Kaitlin: So give me the broad overview of where realtime search engines fit in to the current Web 2.0 ecosystem. What do they offer users?

Tobias: Realtime search helps users discover the signal in the stream – or rather, torrent – of realtime social noise. As the number of users explodes on services like Facebook and Twitter, and as more and more content gets shared between friends on these services, users need a kick-ass search engine to organize all that content and make the good stuff easy to find. That’s what OneRiot does. Search for something as heavyweight as “Goldman Sachs” on OneRiot and we’ll find you the news and stories that are really resonating with people right now as the bank comes under scrutiny from the SEC. Search for something as entertaining as “Lady Gaga” on OneRiot and we’ll find you the pictures and videos that everyone really wants to see right now. All of that is informed by the content that people are sharing with their friends right now on those topics.

Kaitlin: Are realtime search engines and apps going to replace traditional search engines? Or are these veterans adapting well and making their search more “realtime”?

Tobias: The “realtime web” has brought with it the biggest shake up in the Search Industry for a decade. A document-centric approach is being replaced – or at least augmented – with a people-centric view.

Historically, the big search engines were based on an exhaustive index of all the pages on the web. They then determined the relevance of each page and ranked their search results based on factors like the number of in-bound links to that page. This document-centric approach treated the web like a big reference library. Top results were authoritative but static. Search for “Barack Obama” and the top result would be the official White House site or a Wikipedia page for the President. Dependable stuff, but not really reflective of why you would be searching for Obama at that precise point in time.

In the last 18 months, realtime search has radically changed both the users’ expectations of the type of results that a search engine can deliver, and also changed the underlying architecture of how search engines are built. Rather than exhaustively indexing the web based on link structure, realtime search engines like OneRiot index the pages that matter most to people right now. We do this with a people-centric view of the web, based on “social signals” – the content that users are sharing because it’s interesting or important to them at any moment. For example, links that are currently being shared by friends on services like Twitter and MySpace reflect what’s resonating with people right now. At OneRiot we index those pages in realtime. Then we rank our results in order to reflect current social resonance of a particular web page – what we call “the Pulse” of that page. This means looking at realtime social signals, rather than inbound links, to determine how relevant a page is right now. We’ll look at factors like the Velocity of Social Syndication (i.e. how fast is a link being shared right now across the web) and People Authority (i.e. who is sharing a link and how influential are they) to help determine our top result.

Of course, if you want to search for your dentist’s phone number, or find a blog post that was published several years ago, then traditional search is very, very good at that. The library metaphor works well there. But if you want to find out what’s going on with “Barack Obama” right now – what people are saying, what links people are sharing, what’s the buzz on Obama – then a realtime search engine will deliver far superior results. Realtime search will find fresh, socially-relevant search results, linking to stories that are resonating right now. And a lot of people are looking for this type of search results. In our white paper, The Inner Workings of a Realtime Search Engine (http://bit.ly/WhitePaper), we note that perhaps 40% of all search queries display an intent that would best served by results from a realtime search engine. That’s a huge segment of a very big market – so the opportunity is very real.

Kaitlin: What are the current models for advertising on realtime apps and search engines? Could you please describe OneRiot’s plan?

Tobias: Contextual advertising against search results is a very well understood business model. We do that on OneRiot with our system called RiotWise. We also syndicate these ads to monetize other realtime web applications such as Twitter apps (e.g. ÜberTwitter), Desktop apps (e.g. Digsby) as well as search engines.

However, for this to work against realtime search results, or in a realtime app, our system has to deal with two major differences from the traditional search paradigm. First, users on the realtime web are, essentially, trying to find out what’s going on right now for any particular subject. So if an ad is going to be clicked on, it must be useful to the user in that context. Secondly, audience attention on the realtime web can shift radically an unpredictably from minute to minute as new stories emerge and new topics start to trend. To monetize this audience, effective ads need to be in place ahead of these spikes in interest for any topic. RiotWise nails both of the problems.

In terms of content, RiotWise matches users to fresh, relevant content from our paying partners in realtime. RiotWise serves up relevant links to web pages from our content network that helps users find out what’s going right now. For example, if the user is trying to find out what’s going on right now for “iPhone,” RiotWise will serve up paid-for links to freshly published, quality content – news, stories, videos, etc. – concerning the latest developments with Apple’s mobile phone. When these links are relevant to a trending topic, relevant to what people are talking about on the realtime web, the click thru rate (CTR) goes through the roof. In other words, if iPad is trending, and we promote the hottest accessories for the iPad, realtime web users love it. Even on mobile, the CTR goes from an industry average of 0.1-0.2% to over 1% and sometimes even higher. We’ve had CTRs as high as 8% when we really nail it.

To deal with the attention spikes, we leverage our realtime search infrastructure. We index content and products that our advertisers like to promote in realtime as they have it. E.g. Coca-Cola had a campaign that produced content around the NCAA tournament. We had their content in our index the minute it was published, and as it became relevant to trending topics our engine automatically pushed it out across our network in realtime. All ads are created on-the-fly — as the content becomes available — so they are always ready to be matched to our trending topics.

Another key difference is that on OneRiot, advertisers don’t buy keywords. That process would be too slow to monetize on the realtime web. You can’t ask an advertiser to buy a keyword after a term has become trending. The traditional creative ad process doesn’t work in this context – by the time the ad is created, users have already moved on to the next hot topic of the day. Instead, advertisers simply have OneRiot index their content, and we automatically construct ad creative on-the-fly that links to relevant pages on the advertisers’ site, all matched in realtime. We’re always ahead of the curve.

Kaitlin: You’ve probably been asked a lot about Twitter’s monetization plan recently. Put simply, what do you think? How does this match your thoughts on advertising in realtime?

Tobias: We love Promoted Tweets! For one, Twitter is now helping to educate media buyers that they can reach an audience on the realtime web – and that’s good for everyone with an advertising-based business model in the Twitter ecosystem (such as Ad.ly) as well as the wider realtime social web (such as ourselves).

Obviously Twitter’s model is quite different to ours – and the two complement each other nicely.

With Promoted Tweets, a brand like Starbucks can reach Twitter users with a pre-defined message that will be of relevant to those users explicitly interested in Starbucks. Today that means users who are searching for “Starbucks,” or other keywords that Starbucks have bought. Those users might see a sponsored tweet at the top of their search results.

RiotWise is more geared at the wider intent of realtime web users, which is to find out what’s going on right now. Advertisers do not have time to react to trending topics to get their brand in front of the engaged audience at every point in time, so we do that for them leveraging our realtime search infrastructure as noted above. We’ve been in the market with RiotWise now for about 6 months. It’s proven with advertisers(they pay good rates to reach the realtime web audience), it’s proven with the end consumer (the click through rates are very healthy) and it’s proven with distribution partners (who take a rev share). In April we will deliver half a billion RiotWise ads across the realtime web – so this network is growing very fast.

Kaitlin: I thought OneRiot’s blog post on this issue was quite illuminating. From this, what do companies need to know about advertising with you or other realtime engines? What’s the paradigm shift they need to have?

Tobias: If a brand advertiser, or a media company, is already invested in social media marketing, then RiotWise will make perfect sense to them. In that sense, it doesn’t require a paradigm shift, and advertisers should feel quite comfortable. Essentially, RiotWise helps extend that social media marketing effort by reaching highly engaged users across our network of realtime web apps. Yes, we apply some “magic” to make sure that the right message is delivered at the right time. And some customers need a bit of education on why they do not need to buy keywords from us (when they see the results from our realtime matching algorithms, that usually does the trick!). That aside, it’s a very understandable value proposition that delivers great results. The paradigm shift lies in the underlying model and the architecture system itself. That stuff is truly innovative. But advertisers want to know about ROI, not how smart our technology is. We have to deliver on both – and so far so good.

~~

Kaitlin Pike is the Community Manager of Web 2.0 Expo. She can be reached @w2e or @kcpike.

Tobias Peggs is the president of OneRiot and will be on the panel for Monetizing the Refresh – Advertising and the Realtime Web taking place Thursday, 11am in room 2010.

Trackback URI | Comments RSS