Moneyball – highly recommended

Yaron Galai4 Comments

moneyballLast week I participated in a roundtable organized by Carmel Ventures, with probably 20-30 entrepreneurs in the room. I recommended they all read Moneyball by Michael Lewis – a book I read a couple of years ago. I think most of the people in the room didn’t have a chance to take note of the recommendation, so I’m repeating it here.

Moneyball is a must-read for any entrepreneur. It has nothing to do with entrepreneurship (it’s about baseball), and it’s not a very good book (in the literal sense, I mean).

But if you read it with an entrepreneur state of mind, there are great lessons to be learned… 2 have been particularly useful for me:

  1. Look at the same game differently – Most startups compete with many other companies for the same market share. In the online world in particular you are likely to be competing with hugely successful companies like Google, Yahoo, Amazon, etc. You can’t win by playing their game. But you can absolutely win by playing a different game in the same space. An example that comes to mind is Yelp – while all other yellow pages and review sites are focused on getting readers and selling to advertisers, Yelp seems to be playing a completely different game. It looks like they are focused obsessively on their reviewers – giving them tools, recognition, community etc. Yelp is looking at the exact same market as all other yellow page companies are, but playing a totally different game.
  2. Understand what metrics really matter for your business – the metrics that *seem* to matter for your business might not be the ones that really determine your eventual success. Moneyball shows how the Oakland A’s found that most of the metrics that baseball teams have been using for decades (batting average, # of steals, etc) had little or nothing to do with how successful the teams were. Another example is a post I wrote recently about how PV’s and ad revenues might not be the best success metrics for publishers.

Bottom line – highly recommended. Get it here at Amazon or here at Audible.

Share this post!

Pricing Display / Demand Creation Ads

Yaron Galai4 Comments

Following Fred Wilson’s post about display advertising, and my post about demand-creation vs. demand-fulfillment advertising, I had another thought I’ll share here regarding the pricing of “display” advertising (or in my terminology – Demand Creation Advertising)Fred Wilson says:

“…However, this study does not determine what the right price for display advertising is. Search is measurable by virtue of the cost per click and display is not. And if click thru isn’t the right measure, then display has an issue...”


And I agree with Fred 100% on that… While auction-based CPC (cost-per-click) is a great way to price the value of demand-fulfillment-ads[1], it is certainly *not* the right way to price advertising which is introducing a brand and creating demand. That’s because the users are *not* in immediate shopping mode, and the effect of the advertising on them is not necessarily embodied in an immediate click.

So following my thread, I had an idea which is slightly theoretical at this point, but could serve as a way to price Demand Creation Ads in the future:
Search ad inventory is inherently limited. That is why CPC bids on search terms are bidded up over time. In other words, the bids advertisers place reflect the balance between the value they get from each click, and the amount of clicks available on their keywords.

If this is true, then a mechanism that creates a lot more search queries and clicks should increase supply, thus reducing the cost of each keyword. One such mechanism are Demand Creation Ads (or, if you prefer – display ads). These ads don’t necessarily create any immediate clicks, but they should create more searches for the brand down the road.

Assuming all this could be aggregated and tracked efficiently (a big IF), you could then price display ads based on one metric – the increase they create in future searches.
Then the next step in pricing would be to identify a common denominator between the pricing of search ads and the pricing of display ads. So for example, as an advertiser I would know that I could pay $1 per search click, or I could pay 50c for display ads that created enough future searches to reduce my search CPC to 40c per-click.

Again – I’m not sure how doable all this is, but if it were – I think this would bring on the golden age of display/demand-creation ads on the web, making them suddenly as efficient as search ads are perceived today. 

 UPDATE: John Battelle just posted some interesting data from eMarketer on the effect of demand-creation-ads on the # of search queries —————————–>

=======
[1] It can be argued that CPA is a more perfect pricing model than CPC is, because it ties directly to the only real conversion metric that matters – the purchasing of a product. That is technically true, but I think that CPA has a few inherit deficiencies which make it a pretty bad solution for pricing ad inventory:

  1. It moves all the risk from the advertiser to the publisher. CPC perfectly balances the risks on both sides.
  2. CPC is a wonderful common denominator that allows all advertisers to participate in a single, easy-to-understand, marketplace. There is no comparable denominator in CPA.
  3. CPA requires a LOT of data points in order to establish the value of each ad and its probability of generating revenue. That makes it completely ineffective for a centralized and efficient marketplace.
Share this post!

The two types of ads

Yaron GalaiLeave a comment

Fred Wilson has a great post about display advertising discussing a recent comScore white paper about the subject. Fred's bottom line:

"The basic insight from the report is that display advertising does not normally result in an immediate click. That makes sense because the ad is not being presented in a moment of purchase intent, like a search ad is. But the ad does create interest in the product or service which is realized at some later date in the form of a site visit, a search query, and possibly on online or offline purchase."

I think that the common distinction between "display ads" and "search ads" is not the right distinction… The distinction should be between:
  1. Demand Creation Ads – these are ads that notify the consumer about a brand, it's values and reasons to purchase it (whether factual or emotional). Display ads would typically fall into this category. 
  2. Demand Fullflilling Ads – these are ads that offer the consumer an effective shortcut to fullfiling a purchase that s/he has pretty much decided to execute. Search ads would typically fall into this category.   

Demand Fullfilling Ads (DFA's?) is a fairly mature business on the web. Demand is best expressed through a search box, and search ads are an extremely effective way of fullfilling that demand. As comScore noted – the immediacy of conversion on these ads is very high, however in terms of raw volume – there aren't many of these ads. 

However, for Demand Creation Ads (DCA's?) there is no such Google yet. DCA's have to be interesting, not necessarily relevant. That's because the consumer has not yet expressed their intent, and therefore 'relevancy' is no more than an educated guess. 

To date, the DCA's on the web can be categorized into 2 categories:
  1. Interruption ads – banners, etc are intended to capture an audience consuming content and interrupt them with ads. 
  2. Contextual ads – text ads that algorithmically emulate search advertising on non-search pages. 

Contextual ads may be relevant but, they certainly are not interesting. And therefore they mostly fail at being effective demand *creation* ads. They are also not effective as being demand *fullfilment* ads because they are presented to users who have not expressed their commercial intent. 

We're only scratching the surface with Demand Creation Ads. The future lies in the word "interesting", not necessarily in "relevant" and I think there are exciting opportunities to displace and significantly improve on both interruption ads and contextual ads. 
Share this post!