Feb 11

Coincidence or Acute Audience Targeting in Ads

Bargin and Luxury in an Ad pair

Targeting the Rich and Thrifty

Why did these two ads appear side by side as and when they did? Could they possibly have known I recently visited both slickdeals (thrift) and sites containing Mercedes (wealth) content?

Straight out the gate, this was likely coincidence. Disclosure: I work deep in the underbelly of internet advertising, and specifically on the audience segmentation and targeting technology that powers modern day website ad serving. The sheer amount of information exchange and processing required to match this up in-real-time in a cohesive and sale-able way is not (yet) happening.

Google is certainly close. And the more I think about it, the more I think this wasn’t coincidence at all. All the disparate information necessary to make the map is in their arsenal. They could have made this match.

A few factoids feeding the not-coincidence conclusion:

  • I primarily use Chrome
  • I am signed in to a Google account
  • I use Google search often
  • most every blog and site has Google Analytics
  • Ads were served by Google (adsense)

What do you think? Have we entered the era where even remnant banner ads speak specifically to our interests and tenancies?

Jan 10

Commonalities google knows about relationships

boyfriend togirlfriend to

shocker - boys and girls care about different things

Fun idea implemented by predictably irrational, to probe google for common questions from boyfriends and girlfriends. The auto-complete dialog box below the search box shows the most commonly asked questions that fit the criteria typed in so far.

So, I took it a little more generally:

also gives some idea of age of askers

also gives some idea of age of askers

Somewhat surprisingly, swapping guys for boys and women for girls yielded essentially the same suggested query set.

Sexuality type questions top the list; are the days of asking Mom or Dad this stuff gone? Going even further, what if this means we are we raising generic kids? Ones who’s primary environment for learning about themselves is the global interweb? Not real sure this is a fun train of thought anymore.

Dec 09

V8 benchmark suite v5 – my numbers

Google has this benchmark suite thing you’ve probably heard about. Here are my numbers for the browsers I use most often on my machines. First and only run results below:

Windows 7 Ultimate 64-bit, 4GB DDR3 1033MHz, Intel 2.4GHz C2D P8600 Linux Ubuntu 8.10 32-bit, 2GB DDR2 800MHz, Intel 2.2GHz C2D T7500
Opera 10.10
Score: 199
Richards: 151
DeltaBlue: 177
Crypto: 126
RayTrace: 338
EarleyBoyer: 622
RegExp: 86.4
Splay: 203
Opera 10.00
Score: 125
Richards: 87.3
DeltaBlue: 103
Crypto: 69.3
RayTrace: 235
EarleyBoyer: 378
RegExp: 62.9
Splay: 137
Firefox 3.5.6
Score: 383
Richards: 1239
DeltaBlue: 106
Crypto: 746
RayTrace: 261
EarleyBoyer: 355
RegExp: 242
Splay: 554
Firefox 3.0.15
Score: 120
Richards: 90.7
DeltaBlue: 92.2
Crypto: 132
RayTrace: 105
EarleyBoyer: 117
RegExp: 129
Splay: 200
IE 8.0.7600.16385
Score: 98.7
Richards: 50.9
DeltaBlue: 56.3
Crypto: 69.6
RayTrace: 102
EarleyBoyer: 170
RegExp: 124
Splay: 213
IE 6 (Windows XP VM)

Richards: 8.7
DeltaBlue: 3.27
Crypto: 6.69
RayTrace: 16.2
EarleyBoyer: 8.64
RegExp: …
Could not get past 89% completed… ran “forever” on RegExp test
Score: 3754
Richards: 3837
DeltaBlue: 3918
Crypto: 3062
RayTrace: 5351
EarleyBoyer: 6212
RegExp: 1066
Splay: 6445
Score: 3455
Richards: 3272
DeltaBlue: 3523
Crypto: 2863
RayTrace: 4938
EarleyBoyer: 6373
RegExp: 1061
Splay: 5335

No doubt Chrome blows these #’s away update – big time confirmed. What company would release a benchmark suite that didn’t favor its own browser? Benchmarks are an engineers marketing. ┬áMy hunch is that Safari would fare well on these as well.