Dogfood

« October 2008 | Main | December 2008 »

November 25, 2008

Ads in user-uploaded videos – new work from Stanford

Those smarty-smart smart people at Stanford (which gave us Sergey and Larry, Sun, and, ahem, Jerry) have developed some software to make it possible to insert dynamic images or videos inside the environment of another video. I’m not talking about pre-rolls, post-rolls or overlays here; the inserted images/movies (which could of course be ads) are rendered as if they appeared on the surfaces within the “host” movie. You’ll get the idea when you watch this video:

This is pretty clever stuff – akin to what our in-game advertising subsidiary, Massive, does in Xbox 360 games; except with Massive, the in-game geometry is already known, whereas this software figures it out. Now all they need to do is to make the regions clickable, and they’ve got a startup company, right there. Will be snapped up by YouTube before you even finish reading this sentence.

You can try the technology for yourself, called ZunaVision, here (sample videos and images are a bit lame, mind).

del.icio.usdel.icio.us diggDigg RedditReddit StumbleUponStumbleUpon

November 21, 2008

Brandt Dainow gets over-excited again

hs_dainow_brandt After his breathless article last year, proclaiming Google Analytics to be something like a cross between the second coming and Barack Obama, Brandt Dainow seems to have soured on the big G, proclaiming this week that GA contains ‘disturbing inaccuracies’:

Google Analytics is different from other products in that it has been intentionally designed by Google to be inaccurate over and above the normal inaccuracies that are inevitable. These inaccuracies are so glaring that most people are getting a very false picture of what is happening in their sites.

Dainow’s main beef with GA is two-fold:

  • It treats single-page visit as valid visits (i.e. it doesn’t remove them from visit counts or other related measures)
  • It includes single-page visits in average visit duration calculations

He also remarks that Google did in fact change the way that GA calculated average visit duration last year, but then changed the calculation back in the face of user pressure:

Google intentionally rolled Google Analytics back so that it produced an incorrect average duration…It's been that way ever since -- Google is intentionally and knowingly providing inaccurate numbers because a few people preferred neatness to truth.

Brandt then proposes two alternative measures - ‘retained visits’ (the count of visits with more than one page impression) and ‘true average duration’ (the average duration of retained visits). These metrics are not without some merit – it’s useful to know how many visits contained more than one page view, and the average duration of these visits. But Brandt goes on to assert that these two metrics should replace the standard measurements of visits and average duration in GA and (presumably) other tools. This suggestion is ridiculous, for the following reasons:

  • Contrary to Brandt’s assertions, there are a host of scenarios where a single-page visit is a perfectly valid visit, including, for example, this blog, for crying out loud, which has a high proportion of single-page visits because readers either just read the homepage and leave, or click through to an article from their RSS reader. So chucking all these kinds of visits out is crazy.
  • Whilst the inaccuracy of including single-page visits in average visit duration calculations is known to be a problem, removing these visits from the calculation doesn’t yield a magically ‘accurate’ number, it just yields one that is inaccurate in a different way. You still have no idea how long people looked at the final page of their visit for, and with a two-page visit this can introduce a huge potential inaccuracy.
  • Such standard metrics as exist in the web analytics industry are the result of long and arduous wrangling. There are no sacred cows, but you need a really good reason to exchange a simple and easy-to-understand metric for one which is more complex and offers no discernible benefit.

Whilst I can understand Brandt’s motivations for posting these ideas (which, I imagine, lie somewhere on a spectrum between a genuine desire to spark debate and a desire to generate a lot of traffic to his blog, in which regard I am obliging him), his remarks do irk me a bit (can you tell?), principally because he commits the unpardonable sin of absolutism when talking about web analytics, bandying about words like “truth” and “wrong” when really he is just presenting his own preferences.

When, as an industry, we can’t even agree what constitutes a visit, it’s pretty rich to start decrying one tool or another as ‘inaccurate’ simply because it takes an approach to data that you don’t believe in. And besides, as Brandt surely knows, Google Analytics now has the capability (via its custom segmentation) to calculate the metrics he seeks.

Finally, as every half-experienced web practitioner (of whom Brandt seems to have a low opinion also) knows, the key to success in web analytics is to pick your metrics, stick to them, and measure them continuously as you make changes to your site and your marketing, to see what is working. If you’re looking to increase engagement, and have decided that visit duration is a good measure of this (a debatable point, as it happens), then it doesn’t matter whether you include single-page visits in your duration calculation – if your visit durations are going up, you’re happy. And if your visit durations suddenly jump because your web analytics vendor has changed the way they calculate the metric, this could in fact cause more pain than benefit, perhaps causing you to go to said vendor and say, “Oi! Change it back to how it was!”.

So feel free to read the article, but be warned: it’s not very accurate.

del.icio.usdel.icio.us diggDigg RedditReddit StumbleUponStumbleUpon

November 13, 2008

Online Ad Business 101, Part VI – Ad Exchanges

Yes, it’s time for another Online Ad Business 101 post. This post deals with one of the players that I left out of my first post about the online advertising value chain: Ad Exchanges. If you don’t know what an ad exchange is, now’s the time to learn, since these little-known companies (most of which are now owned by some extremely well-known companies, such as Microsoft, Yahoo and Google) are set to have a major impact on the way the industry works over the next few years. Our own AdECN recently announced the launch of its new “federated” ad exchange, making this post especially timely (in truth, the timing is no coincidence – the announcement was the gentle kick I needed to get this post out the door).

 

The problem with ad networks

Ad networks are great. They must be – there are enough of ‘em (300 the last time I counted). As you’ll know from reading my post about them, ad networks make money by creating a marketplace between publishers (providers of ad inventory) and advertisers (consumers of ad inventory). By connecting many publishers to many advertisers, they create some efficiency and add some value into the bargain (by providing targeting capabilities, for example). Networks sit at the center of their own little web of advertisers and publishers (apologies to Right Media for, ahem, borrowing their little people graphics for this post):

image

The key challenge with being an ad network is that you have to grow the supply side of your business (the publishers) in parallel with the demand side (the advertisers) – there’s no point signing up a huge batch of new publishers if you’ve no one to sell their inventory to. But doing this in practice is extremely hard.

To solve this problem, the ad networks have brokered relationships with one another over the years so that, if a network has an impression that it needs to sell, but doesn’t have an advertiser to sell it to, it can sell that impression to another network. Similarly, in the reverse case, if a network has an opportunity to sell an ad, but doesn’t have the inventory to fulfill the sale, it can buy the inventory from another network. So the picture (with each network’s advertisers and publishers collapsed to one of each) looks like this:

image

In the diagram above, Network 2 can sell Publisher 2’s inventory to Network 1, who sells it on to Advertiser 1. Similarly, Network 2 can buy inventory from Publisher 1 (via Network 1) and sell it to Advertiser 2. In the real world of ad delivery, the ad call is redirected from the publisher to Network 1, and then to Network 2, before finally being redirected to the advertiser’s ad server.

If you add another ad network to the mix, then each ad network can forge relationships with the other two, and trade impressions in much the same way:

image

If there were just two or three ad networks in the world, this might not be a problem. But of course there aren’t – there are three hundred. But each ad network can’t have a relationship with every other ad network; each network would have to maintain 299 relationships, which comes to (299 + 298 + … + 2 + 1) = 44,850 relationships!

So instead, the networks form a kind of ‘daisy chain’ – each network passes off some portion of its inventory to one or more others, which in turn pass some of this off to their own partner networks, and so on. So a single ad impression can pass through half a dozen (or more) networks before finally being fulfilled:

image

Of course, the diagram above dramatically over-simplifies the picture; each of the networks in the chain will have multiple relationships with other networks, and so inventory can take a series of routes from a particular publisher to an advertiser.

This daisy-chaining sucks, big time, for the following reasons (and others):

  • Each network has to maintain multiple bilateral arrangements with other networks, sucking up time and technical resource
  • The more networks there are in the chain, the longer the ad takes to serve
  • Each network wants to take a cut of the cost of the inventory,cutting into the publishers’ margins
  • In addition to the margin problem, it’s very difficult for publishers to get the best price for their inventory, since each network in the chain has to make the best guess around the network it thinks will deliver the best price
  • The publisher has very little or no control over the quality of the ad that ends up being displayed; it’s very easy to insert poor-quality or even malicious ads into the system
  • If the ad is clicked, the click path is very convoluted, and is liable to hijacking (another security vulnerability)
  • The system is completely opaque to the publisher (no network can provide a comprehensive list of what actual ads they served, or had a hand in serving, on a publisher’s site)

 

Enter the Ad Exchange [fanfare]

By now I can almost hear you crying, “But surely there must be a better way!” Well, you’ll be glad to know, there is. Rather than the ad networks all dealing with each other directly, we need some kind of impartial intermediary which can act as a central hub through which the networks can trade. An ad exchange, if you will. Of course, exchanges (especially commodity exchanges) have been around for a long time – as I noted recently on this blog, everything from pork bellies to weather futures are traded today on exchanges around the world today.

The Chicago Mercantile Exchange (one of the biggest exchanges in the world) is even launching an exchange for credit default swaps – those bad-boy financial instruments the opaque trading of which (in a manner which is alarmingly reminiscent of the ad network relationships above) have had such a hand in getting us into the mess we’re all in now.

So if even something as evanescent as a CDS can be traded on an exchange, why not ad inventory? Well, it turns out there are some fairly interesting technical challenges, since the volume of transactions is extremely high and transactions must be completed within a few milliseconds; but those are surmountable. By adding an ad exchange into the picture, the trading relationships look as follows:

image

Now each ad network has just one trading relationship – with the exchange. So if there are 300 networks, there are 300 relationships, and every network is just one ‘hop’ away from every other network.

What this is means is that for a given ad impression on a publisher site, the network that owns that impression can say to the exchange, “what am I bid on this impression (one careful owner, full service history, nice neighborhood, good references, etc)?”. The exchange can then hawk that impression to all the other networks and solicit bids. Depending on the data that is attached to the impression (or a cookie that one or more of the other networks may recognize and be able to attach data to), the various networks may be able to sell that impression for a greater or lesser amount. So the bids come in, the winning bid is selected, and passed back to the originating network; and if that bid is better than what the network could get from its own advertisers, it wins, and the ad is served.

Crucially, there are only ever two networks (plus the exchange) in this transaction. So each network will take a cut of the impression price, and the exchange will charge a flat transaction fee (this is essential to maintain the exchange’s impartiality – taking a cut would introduce bias). Just having two networks in the transaction means more money for the networks and the publisher, and possibly better pricing for the advertiser. So everyone wins.

 

There’s more…

The benefits of moving to an exchange-centric model for ad inventory trading don’t end there. As I said at the beginning of this post, one of the irritating things about running an ad network is having to match demand to supply – as networks grow, they have to recruit both advertisers and publishers. The network model allows one-sided participants to flourish, dramatically increasing the range of ways in which businesses can participate in this market.

image

In the above example, Network 2 doesn’t actually source any inventory direct from publishers – it gets it all from the network, and focuses on being great at selling that inventory to advertisers. Another (perhaps better) name for the kind of company that does this is a Media Agency. Havas has just announced its intention to do something similar to this. You could also easily imagine the likes of Amazon and eBay – both of which have huge rosters of small advertisers, but no corresponding publisher base – to participate in this fashion.

Similarly, Network 3 above has decided to do away with its advertiser customer base and just sell all its inventory to the exchange. In this sense it becomes a bit more like an ad sales house or publisher aggregator than a true network. A company in this mold might be Six Apart, creators of the TypePad platform, which has lots of publisher relationships (including with me), but no advertiser relationships to speak of.

And there’s a third scenario which is even more interesting, which is that the exchange model makes it possible to add value (and make money) without trading any inventory at all. A company like Nielsen might choose to sell the data it has on internet users to the exchange, helping to drive up inventory value, and taking a cut of transactions that use its data.

Building a true ad exchange is non-trivial, mind you, which is why the most significant efforts in this space are courtesy of the Big Three of Google (most visibly via the DoubleClick Advertising Exchange), Microsoft (with AdECN) and Yahoo (via the RightMedia Exchange). And a big question-mark still hangs over how exchanges will earn money – the revenue model is well understood (transaction fees), but whether those fees will be enough to support the exchange’s costs remains to be seen. That’s why it’s the companies named above which are most active in this space, because they all have ad networks of their own that they want to add value and liquidity to, so that they can recruit more advertisers and publishers, and ultimately take over the world (bwwahahaha).

 

Impact

Ad exchanges are poised to have a transformative effect on the online advertising business. Given the current economic climate, you can probably expect these creatures to fly under most people’s radar for the time being – probably until late 2009, I’d say – but their influence will be felt as advertisers find it easier to reach the audience they need (via greater liquidity in the marketplace), and publishers are able to hang onto a bigger chunk of the price of their inventory.

In the latter case, Exchanges could end up changing the balance of power between direct-sold and network-sold inventory – if a publisher can get a better margin (taking into account sales costs) by sending some inventory that was direct-sold to the exchange, they will do so.

But what about networks? They will likely see better margins by going through an exchange for inventory they can’t clear themselves; but exchanges will level the playing field in terms of inventory access, meaning that networks will have to ad value  over and above simple aggregation in order to survive. Competition for publishers may intensify since, if networks are backed by exchanges, there’s less incentive for publishers to have deals with multiple networks. Certainly we can look forward to lots of change – consolidation, specialization, fragmentation – in this industry in the years to come.

For more reading on this from someone who knows much more about it than I do, I’d recommend these two excellent posts on the topic from Mike Nolet, formerly of Right Media.

 

Index of all Online Ad Business 101 posts

del.icio.usdel.icio.us diggDigg RedditReddit StumbleUponStumbleUpon

November 07, 2008

Applied Insights falls into the gaping Foviance maw

neilmason Ok, it’s not quite Omniture acquiring TouchClarity, but I was delighted to read yesterday that my old pals at Foviance have secured the services of none other than Neil Mason through the acquisition of his company, Applied Insights. Neil (whose ClickZ column you should read) has been flying the flag for web analytics – especially from a marketing-effectiveness point of view – for many years, and he’ll be a great asset to the Foviance team. Congratulations Neil and Paul. Neil’s partner and co-founder of Applied Insights, John McConnell, will leave to pursue independent interests.

The only note of sadness for me in this announcement is that the background to the acquisition is the change in the focus of Foviance’s web analytics efforts away from a WebAbacus-oriented technology/consulting solution to a services-only offering based around Omniture, WebTrends and the like. Of course, this is precisely the right thing for Foviance to do, since the web analytics market is now firmly consolidating around the major players; but I’m sad because WebAbacus (which I spent so many years on) isn’t one of them. But I like to think that its influence in the industry lives on.

del.icio.usdel.icio.us diggDigg RedditReddit StumbleUponStumbleUpon

November 06, 2008

In Excel hell?

This will cheer you up – the world’s first Excel-based music video (I kid you not):

[Via adfreak]

del.icio.usdel.icio.us diggDigg RedditReddit StumbleUponStumbleUpon

November 05, 2008

Oh yes…

xkcd_election 

From xkcd. How true.

del.icio.usdel.icio.us diggDigg RedditReddit StumbleUponStumbleUpon

November 03, 2008

Election night: Make sure you get home early

My old friend and co-émigré Bruce Nash has put together a great presentation about how tomorrow’s election will unfold, in terms of when the TV networks will call each of the states. From Bruce’s model, it looks like it’ll be all over bar the shouting by around 8pm Pacific – so if you’re on the West coast, best get in front of a TV set early if you don’t want to miss all the action. You can view the presentation below (jump to around 6:30 for the meat):

The only beef I have with Bruce’s predictions is that he models the poll-to-result swing (the difference between the final poll result and the actual result) for each state using the values from the 2004 election – and there are many reasons (new voter registrations, a different profile of voter turnout, the Bradley Effect) why the swing could be different this time around (to be fair, Bruce does acknowledge this).

But since Bruce’s model predicts that Iowa, Pennsylvania and Virginia will lock up for Obama pretty early on, leaving McCain with very few options, it’s hard to imagine that the outcome won’t be fairly clear by the time folks on the East Coast (and my three-year-old daughter) need to go to bed. In fact, it could even be the case that the result is pretty much known before the polls actually close on the West Coast. Kinda crazy, if you ask me ;-)

del.icio.usdel.icio.us diggDigg RedditReddit StumbleUponStumbleUpon

Subscribe

Enter your email address:

Delivered by FeedBurner