Keywords and click rates are their bread and butter. "We are trying to understand the mechanisms behind the metrics," says Qing Wu, one of Varian's minions. His specialty is forecasting, so now he predicts patterns of queries based on the season, the climate, international holidays, even the time of day. "We have temperature data, weather data, and queries data, so we can do correlation and statistical modeling," Wu says. The results all feed into Google's backend system, helping advertisers devise more-efficient campaigns.
An interesting article from Wired about Google and how it sees value in its data.
Hat tip to ogletree
those who adopt data driven decision making are screaming ahead
All the banks, investment bankers, fund managers, economists, policymakers, government agencies, etc. were . . just swagging it, eh? That's why the world economy is in a meltdown? Due to a lack of data driven decision making?
Darn those swagging credit card holders and their lack of data driven decision making!
In complex situations, data is a also way too subjective: what do I measure A, or B (or C, but I didn't know C existed).
Data handling is not like pure math or physics (well defined problem, equations, sometimes closed form expressions), it's more statistical and actuarial, and thus imperfect.
Nonetheless, there is vast amounts of knowledge to be unlocked with all the data spewing tools that Google has put in the Webmaster's hands: Analytics to track how your site is perceived by G, AdSense channels to see what works are my 2 most important.
All the banks, investment bankers, fund managers, economists, policymakers, government agencies, etc. were . . just swagging it, eh? That's why the world economy is in a meltdown? Due to a lack of data driven decision making?Darn those swagging credit card holders and their lack of data driven decision making!
i meant in internet marketing, but you bring up an interesting point about the world, beyond webmastering for money.
while you may question whether the point i made is the reason for these things, you do so emotionally and without basis. i didn't make any comments about world economies or meltdowns, and the report i was commenting on was clearly defined and restricted to our usual topic fare here. so why the emotional extrapolation from you? please do let us know how you made that decision and what you hope to accomplish with it.
[edited by: RhinoFish at 2:01 pm (utc) on May 26, 2009]
Very true and very valid point, but I didn't say to begin haphazardly collecting any data you can locate and start making decisions from it. Being data driven includes refining and optimizing data collection, not just analysis and beyond. It also includes continuous exploration of new tactics and techniques to uncover, and find utility in, new data.
[edited by: RhinoFish at 2:00 pm (utc) on May 26, 2009]
Another valid, true and important point! I didn't say being a data driven marketer is perfect, I said it's gives you an advantage over others because it's better than not being data driven. So it's not a question of finding perfection, it's optimization, on a relative scale.
All the banks, investment bankers, fund managers
economists, policymakers, government agencies
The whole mess was bound to end in tears ..some of us said so years ago even here ( threads about affs whose main business was providing leads for debt consolidation or sub prime providers via their aff "schools" ) ..and the worst is yet to come when the insurance policies on the reselling of toxic assets that went on before 2008 have to be paid ..lehmans and madoff will seem tiny and inconsequential when that financial tsunami rolls over ..or we'll have a war or two to distract the gallery and reduce the mouths to feed ..plus que ça change ..etc.
The data is/has always been available ..collating and interpreting it is the easy part ..facing upto and acting upon what it tells one is the hard part that needs intelligence "guts", drive, resolve, and the ability to think further than next week or quick bucks..
Google given data is however different ..it's source ( Google ) is not impartial ..it has it's own aims and agendas ..they may or may not coincide with mine or yours ..so IMO their pronouncements are suspect and never to be trusted 100% ..they will help you to help them better ..but they are always their first priority.
Reminds me of the story of a group of blind men trying to discern the nature of an elephant by touching it ..one has a leg ..to him it is a like a tree ..another has the tail ..to him it is like a donkey ..one has an ear ..to him it is like a great bats wing ..and one has the trunk and to him it seems as a great snake ..
If they all bring food it will allow itself to be touched ..
But if it sees tastier food a little way off it will move on..and elephants do not step around what they consider to be ants ..they step on them ..not from malice ..they are just individually too small to be considered ..in an elephants scheme of things ..
Here we are exchanging data and our personal perceptions of the elephant that is Google ..here is a very usefull place to try to pool our experiences and knowledge of the animal..as long as we remember not to believe totally anything that the elephant may have to say or try to hint to us on the subject ..
To Google ..we are all expendable ..there'll be another webmaster, advertiser, publisher along on a minute ..and another ..and another ..and that is the way of things and is how it will be ..
I personally dont believe that the 'economy is in a meltdown', but is in a big state of change (actually partly due to the internet), and that isnt being measured correctly.
Thing in the marketing world is so many still don't collect data or look at it, don't do anything with it or discount it and still shoot from the hip. Data --> Information --> Knowledge -->Action --> Analyze --> Repeat.
In most cases, those who do, particularly on the web will be crushed by those who pay attention to what is going on under the hood and let that be their guide along with some common sense gut checks.
Exactly! And I really like the way you phrased this. One big problem is ignoring all data. Another problem, also huge in its own right, is not applying common sense in gathering or analyzing data. The latter causes more ROI pain, in my experience. Blindly following certain numbers can get you in some real serious ROI trouble.
If success really does come down to a matter of prowess in terms of analyzing data, it seems that the conversion optimizer would be pretty stiff competition...
...at least in theory--I'm not very up to date on current opinion of the optimizer. I use it on one campaign and it hasn't put me out of a job yet.
that said, if you're not doing any analysis, then allowing the conversion optimizer to make choices for you, would likely improve things. then again, if you're the type not doing any analysis, i doubt you'd fiddle with the conversion optimizer either.
Thus making the conversion optimizer just a tool at our disposal rather than a robotic replacement, as I originally phrased it.
Companies need to understand the sleight of hand pulled in reporting media (banner) campaigns when lumping "view through" conversions in with click-through conversions. If the measurements changed to be the same as search where someone must click on an ad and buy something, not just get cookied because they were on a page where there was an ad, CPMs in media would plummet to nothing. Ad networks and big media agencies will do anything in their power to keep the measurements from changing.
Data is very useful in PPC campaigns but paid search is fundamentally a language problem and not so much a bid problem (where everyone puts all their efforts and millions of dollars). Get the language right and the bid stuff gets a lot less important.
You need to collect the right data, make all decisions on the web under the constraint of "will this help us make more profit now and/or in the future?".
It takes data and a "systems" thought process - see Peter Senge in the "5th Discipline" (book) and Goldratt, "The Goal" (theory of constraints). Both were penned well before the web got big but outline the type of non-linear thinking required to succeed on the web.
Brute statistical analysis doesn't work so well on the web, IMHO, but it seems that that is what the engines use to analyze data. Maybe it is necessary with the volume that they have. Pure statistical analysis is no substitute for thinking about the intent behind the action and looking at the actions (data) on a pretty granular level to get a better "feel" for the intent behind the query or action.
If they really believe that the loose match types are good and that an algo that matches out terms does anything more than drive click prices higher and generate more revenue for the engines, then I think they are kidding themselves.
It will be interesting to see if Facebook can do more with their data than ad networks. Both can collect all kinds of stats about people and their behavior online. Person X visits these 5 sites, did a search for these terms over the last 4 weeks, is 45 years old, etc... and they can serve an ad to very targeted demos and profiles. They cannot take intent into that equation and I think intent is the single most important thing. With search, there is definitely intent because the person is searching for something and intends to find something. By looking at the stats from search you can get a pretty good idea of whether or not you are anticipating intent of the keywords accurately.
Its ironic that with all our computational math re SEO blah blah blah. Matt Cutts puts it so simply in his recent interview. Its compelling to bear in mind that this guy kicked BMW off of the Internet. He espouses that data is indeed to key to profit - not Black Hatting
[youtube.com...]
