Recently, I’ve found myself in a series of meetings in both the Silicon Valley and in Israel with technologists, media people, and venture capitalists – all of whom seem to be equally despondent about the state of the online advertising start-up. The online advertising industry is in the final stages of consolidation, they argue, technology doesn’t matter all that much, and opportunities for online advertising start-ups to break into the revenue stream are going to be scant to non-existent. What’s the source of all this pessimism, and where might opportunities exist for venture investors in this market?
Let me start by saying that the online advertising market is extremely complex (probably far too complex for it’s own good). A good friend once gestured at the floor of the London Ad:Tech conference. “You see all these companies?” he asked with a smile. “80% of the clicks on the Internet pass through 80% of these companies. It’s a total mess.” He was exaggerating of course, but it’s not far from the truth. To help make some sense of this, let me put forward a few observations that I’ve gathered and try to tie them together into a thesis. As always, thoughts, comments, and critiques are welcome.
- Consolidation in the industry is very real. There has been a trend of acquisitions large and small over the past several years. Traffic is being concentrated in fewer and fewer major media companies. The buy-side is consolidating as the major agencies snap up digital agencies and lock up market share. In addition, networks and exchanges are consolidating as the bigger players see the value they can extract from scale. As the market for online ads becomes more efficient, the need and opportunity for niche players has contracted dramatically. This trend is raising the bar for start-ups. To achieve significance, a start-up must penetrate one or more of a handful of major players in the industry – and that can only be done if the offering is very valuable and very unique.
- The evolution from ad networks to ad exchanges is a profound shift, the implications of which are only beginning to be understood. Ad networks aggregate buyers (agencies) and sellers (publishers) with the promise of superior targeting and easier transactions. Their profit margins, historically, have been based on the spread they were able to achieve by buying low and selling high. The thing is – publishers prefer to sell high and advertisers prefer to buy low. Ad networks will continue to survive in certain markets and niches, but for the most part, the market has evolved to the far more efficiency ad exchange model. An ad exchange is exactly what it sounds like: a fundamentally neutral marketplace where buyers ask and sellers bid and transactions occur where there is a match. The ad exchange model is built on high volumes and low margins – they don’t have much of a spread to profit on and make their revenues from a relatively low commission on every trade. Many networks have tried to adopt aspects of the exchange model, but as profit margins contract, the exchange model naturally evolves towards scale – and we are seeing that today. There are already several major exchanges (Google’s Ad Exchange, Yahoo’s Right Media, and Microsoft’s AdECN to name the major ones), and there is no reason to expect more than a handful to persist over time.
- Automated online advertising will become become increasingly liquid. This is already happening with the shift towards real-time bidding. As they compete for volume and share, exchanges are going to make it easier and easier for advertisers and publishers to transact in real time. Specific audiences, contexts, keywords, cookies – it is all going to be up for grabs in real time (if one wants).
- Direct sales is still king of the castle. For any publisher of any size, direct sales remain the best way to sell quality inventory. Advertising will remain a relationship business and technology tools that help publishers to understand their audiences and validate the impact of campaigns will be weapons in the arsenals of direct salespeople who will continue to drive the bulk of revenues. The question is what will happen to “remnant” inventory (the stuff the direct sales guys can’t sell) and inventory of publishers that do not have direct sales teams.
- Targeting is a cocktail – and those cocktails are getting more complex. As a venture capitalist, I’ve had the privilege of meeting with many start-ups each with their own targeting technology and approach: contextual, semantic, retargeting, demographic, psychographic, dynamic content, etc. Unfortunately, it’s getting harder and harder to get excited by companies that offer a one-dimensional improvement in targeting technology. Those companies may be able to raise the value of impressions all things being equal – but all things are not equal. As exchanges drive the market towards ever more dynamic advertising transactions, the winning solutions are going to be able to investigate multiple targeting possibilities on the fly. The bigger the publisher (or the advertiser or the network), the more data is going to be available to drive decisions across the daisy chain and within it – and the better the clearing process is going to be in the matching of audiences and inventory. Point solutions may have a place, but my guess is that they will be increasingly relegated to the role of market participants – buying and selling remnant inventory at a profit based on their proprietary algorithms. If a company has really powerful algorithms, it might be able to make good money – but these companies are unlikely to be good venture investments. Like algorithmic hedge funds, the smart online advertising companies will raise a small amount of seed capital and trade for their own book.
- The locus of advertising intelligence will shift. With the shift from networks to exchanges, the locus of “intelligence” will shift from the center of the market to the extremes. Whereas once publishers and advertisers relied on the intelligence, experience, and good will of the networks to match supply and demand, the shift of volume to exchanges means that the intelligence driving pricing and purchasing decisions will migrate to the two ends of the spectrum. We have already seen this happen. On the publisher side (the sell side), yield management companies such as Rubicon Project, Pubmatic, Admeld, and others offer publishers the ability to optimize the way they offer their (most remnant) inventory across a spectrum of networks and exchanges, each “competing” for that inventory on the open market. Likewise, companies such as MediaMath and X+1 have emerged that offer similar yield management solutions for the buy side.
- Format companies face an uphill battle. Companies offering creative ad formats face an uphill battle in this new, streamlined online ad universe. This is for two principal reasons. First, most ad inventory purchases are made by checking off boxes on the “Chinese menu” of fairly standard choices. In order to get on that menu, a new format needs to achieve some level of standardization (ideally through the IAB). This is hard to do for a new format, and the process of wrangling buyers and sellers together around a new format can choke a start-up. Secondly, online advertising (at least the profitable kind that VCs like) is all about scale, and scale is all about reach. Unless publishers are ready to sell your format, you cant bring advertisers on board no matter how hard to try – and the revenues of a format company will always be bounded by the reach of the format.
- Data is tradable too. Companies such as BlueKai and the Israeli company Exelate have shown that data on user behavior is just another commodity that can be moved readily from place to place to enhance targeting. This further levels the playing field between publishers and networks – and increasing the ability of advertisers to buy specific audiences. Increasingly, online advertising is less about impressions and context and more about audiences. Context plays a critical role, but now that behavioral data can be moved around easily to be leveraged far from where it was generated, context is just part of the picture.
All of this means that the online advertising industry is changing and consolidating. That said, I think the generalization that “there is no more room for online advertising start-ups” is only as true as most generalizations: It’s mostly true – but its sometimes wrong….and as a VC, I’m looking for those companies that can find and exploit islands of opportunity in the new era of online advertising. Below are some areas where I’m looking, but I’d be happy to hear suggestions for more:
- Yield management. It’s a tough market, and a crowded one – but I think there is quite a bit of evidence that the yield management space has yet to fully mature. Developments such as ad exchanges, video ads, real-time bidding, new sources of user data, and opportunities for dynamic creative create tremendous complexities for publishers, networks, agencies, and advertisers. Existing yield management tools were, for the most part, born in a simpler era and are not up the task. The good news here is that there are lots of potential customers (on both the buy side and the sell side) and that the exchange-based eco-system is ripe for optimization – in fact, it demands it. A new Israeli start-up, IgniteAd, is one of the more interesting upstarts on the publisher side – but they face a tough battle against incumbents.
- Video and video analytics. It’s become clearer that online video is here to stay and that a revenue model will emerge, but we’re not there yet. Video brings a new set of challenges in terms of analytics and profiling – and there may still be room for a start-up to emerge that will help organize this still-nascent space. Video (or other interactive formats) also hold the promise of reversing the phenomenon of analog ad dollars turning into digital pennies. Interactive formats could provide the emotional resonance that will allow branded campaigns to achieve online the same level of impact that offline adds achieve. Innovid, an Israeli start-up with unique capabilities around interactive video advertising and analytics, is one of the emerging leaders in this space (and a Genesis Partners portfolio company).
- Hyper-dynamic advertising, With more and more data on user behavior available and an increasingly real-time environment, we are likely to see more and more hyper-dynamic advertising. By hyper-dynamic, I mean that all elements of the ad itself (creative, text, location, format) will increasingly be customized on the fly to drive performance. The Israeli company Dapper is one of the leaders in this space, and an important company to watch.
- Advanced user profiling and psychographic analytics. Ask the large social networks and they will claim that they are already doing this, but I believe opportunity still remains for powerful semantic technologies and profiling engines to help websites categorize their users better. Advertisers are increasingly buying specific audiences as opposed to impressions – and publishers should be able to provide a much better picture of their users than they are today. Any website that allows users to express themselves (and there are more and more of those) as well as any website with deep and highly specific content should be able to generate a psychographic profile of a user – much like the detailed psychographics that underpinned the offline publishing industry. Israeli start-up Nuconomy, which was just rolled into LivePerson, had some very powerful user profiling technology, although it’s not clear where that technology is headed today. Facebook is experimenting with this, as are start-ups such as Peerset.
If you are working on a start-up that has a strategy to win in the brave new world of online advertising, I’d love to hear from you.