Google’s search share went from 79% to 75%– why that doesn’t matter

So much quibbling among the SEOs of how Bing is stealing traffic from Google or how Yahoo! isn’t what it once was.
Ten years ago, I worked at Yahoo!. And while I’d be tempted to chime in on this tempest in a teapot, they’re missing the point.

That’s me with Jerry and David, who co-founded Yahoo!

People are increasingly spending their time in apps and in closed gardens.
Facebook kicked out Bing, which powered their search, in favor of their own search.
Do you think Amazon is going to let Google ever power their search results?
How about letting Google into your very personal Snapchat experience?

The search game as we know it is over– Google has won with a dominant monopoly position.
The old-timers will argue that Yandex’s 55% share in Russia or Baidu in China at 56% means there’s still competition.
Or maybe they’ll point at how search is still strong and getting more personalized.

But what they miss is that personalization technology, no matter how awesome, is only as good as the amount and quality of the underlying data.
The social network and app economy has far more data than the search engine does– they have no incentive to share it.

In the long-run, whoever has the most robust data about users will do a better job personalizing.

That creates ad revenue, which allows the network to invest in more features (things like gmail or driving cars) and better ad tech.

It’s not that Google is going to be dead or that search is going the way of the print yellow pages.
It’s that search has reached maturity, so Google is having to do things like invest in wearables and Uber.
It’s less about buggy whips and more about Innovator’s Dilemma.

So don’t be like the old geezers sitting on the front porch swapping war stories.

Make sure you know who your customers are, independent of the keywords they search, social networks they hang out on, apps they use, videos they play, brands they like, or places they visit.
The sites/apps/things that they frequent will be the ones providing self-serve ad interfaces allowing you to target exactly these people.

Right now, Facebook has the most profile information and user activity, so they’re the current winner in targeted personalization.
But you know that Twitter, LinkedIn, Pinterest, and even Apple have released their own targeted ad platforms, fueled by their user behavior.

A keyword is not a user

The search engine doesn’t remember what you’ve done last time– you get the same answer each time.
When you set up your marketing automation properly, you’re able to snipe the demand well before it ever becomes a search.

So when we start to see search decline, you’ll know why– it’s like an ounce of prevention versus a pound of cure.
And the cost to both acquire and keep customers via inbound marketing is much cheaper than paying full price at the last minute when the customer finally needs it.

10 Commandments of Running an Ad Network

If you run an ad network, see how many of these 10 Commandments you actually deliver upon versus just say you do:

1.  Thou shalt deliver publishers better eCPM’s: Not the highest payouts, but just better than the competition. That means some ads don’t get to run.  Though we are a marketplace, we want to deliver consistent earnings, which may require us to adjust margins and allocate high performing offers unevenly to key publishers.

2.  All users are not created equal: Focus on the top 10 advertisers and top 10 publishers– not the 700+ other guys who have random thoughts and will waste our time.  Total self-service for them. Tier accounts based on value to the network, which means better performance and personalized service.

3.  Don’t reinvent the wheel: There are common standards for most of what we do (sign-up screens, payment terms, clickfraud detection, reporting, account management, ratecard discounting, etc…).  Copy AdSense and AdWords.  Over-invest where we are different.

4.  Test, test, test: Got an assumption?  It’s probably wrong, so run a quick test and make decisions based on the data. Establish a testing framework so we can do zillions of tests with limited effort.

5.  Payment up-front: We don’t give out loans, ever.  You gotta pay to play.

6.  We have 31 flavors: Don’t argue with them– if they think buying clicks at 10 for a dollar is cheaper than 10 cents each, fine. If they think managing to CPC or CPI or another metric is better, who are we to argue?  If they want to manage their account via, or through their API, that’s great.  Provide all mechanisms– see Commandment #3.

7.  # Follow the money: Let the market tell us what is working and continue to do more of it.

8.  We are not religious:  Let publishers decide what they want to accept. We allow any offer to be on the network, provided it is legal and would be acceptable for a public company.

9.  Everything can be measured: If it doesn’t produce measurable margin, then it shouldn’t be done.  All people, projects, features, advertisers,  publishers, and tasks can be boiled down to their margin contribution.

10.  Transparency: We will show advertisers and publishers their own performance in stunning detail, but not that of others.  All performance questions can be answered by a finite set of automated analytics.

Google Enters the Ad Serving Game Today

Today, Google announced their integration of Google Ad Manager and DART. Users of Google Ad Manager are “upgraded” to DFP for Small Business.  It’s free, as are most all of Google’s products, provided you have less than 90 million impressions a month. If you’re bigger than that, then you have to pay support on the premium product.

So if you’re in the business of selling ad server software, I believe your goose is cooked.  Your prospects are about as good as Netscape trying to charge you $89 a year for their browser, Omniture charging $250k a year to use their analytics software, or AOL charging you $20 a month to keep your email address. There may still be folks who will buy the black magic of ad serving, but it’s pretty clear that few firms can compete with Google on ad optimization, as arguably nobody is better at this when it comes to targeting, eCPM maximization, forecasting CTR, frequency capping, and so forth.

But there still are niches that Google doesn’t play in– or won’t play in.  One such example is social ads that inject personalized information into ad templates.  So while Google has potentially won the game of deciding which is the best ad to show to what person, they have not  gone down to the level of using social data to create personalized messages– so customized that they might have your name, friend’s images, and other details in them.

Look out for social ad serving companies that go beyond retargeting and BT networks. In the same way that there is unintended personalization on Facebook– expect that to spread to the general web.  And once the wave of spammers have had their fill, expect local businesses to run smart local ads using social data.

As common practice, Google also released a few videos with this product update. Below is one for the DoubleClick For Publishers on trafficking in the UI.  Notice that related videos are “sex trafficking in Cambodia”, “human trafficking”, and the like.  Perhaps Google DOES have a ways to go with their targeting capabilities.