Profile cover photo
Profile photo
Contextually (semantic) targeted data for optimal and safe digital advertising
Contextually (semantic) targeted data for optimal and safe digital advertising


Post has attachment
Ad Week in NYC is starting to ramp it's messaging and since we are one of the partners this year they asked us to blog.  It came out this morning.  Lots good. 
Add a comment...

Post has attachment
Big Data Modeling of Semantic Advertising Effectiveness

In my July 17 post I wrote about some work we are doing in the big data arena to model the effectiveness of semantic vs keyword based ad targeting.  

Now some of the results are flowing in.  To date we have completed 5 major newspapers and 2 magazines.  Here is what we find:  a fully semantic approach (no statistical guesses) on the meaning of words averages 54% better reach and 34% better lift when compared to a keyword targeting approach.  This includes such venerable papers as The New York Times, Chicago Tribune and Washington Post.     

Later I will write a longer piece about the top 10 newspapers, top 10 digital magazines, etc.  But I expect the results to stay in this range with a couple of exceptions.  Given a full dump of the ads running on a domain rather than    the online sampling we can observe will make the reach increase smaller but the lift increase larger.  Why?  Because a full sample of ads by its nature indicates better reach to begin with while a semantic approach gives much better segmentation with which to find the right slot for each ad so lift increases. 

The other exception is the use of emotions in targeting the ads.  When I sample online ads and think about the concepts they represent I am get results around the averages stated above.  But in the last run I slowed down to think about the kind of emotions the advertiser likely was trying to convey.  When I use these to target semantically both the reach and lift increase over keyword targeting seem to jump dramatically - to the tune of ~70% increases in reach and lift.  At ADmantX this use of emotions to target ads contextually is something we invented and are happy to report such increases.

It is also satisfying to note that some have recognized the opportunity for ad technology to address "the other half of the problem" - precisely while ADmantX exists and the kind of problems we solve - as noted in the linked piece and I quote:

"For all the innovation in the ad-tech space over the last decade, it's fairly impressive that very few of the core problems of a publisher have been solved. At the end of the day, 60-80 percent of the revenue that publishers bring in comes from their premium inventory, sold on a guaranteed basis -- which represents generally less than half of all their available inventory. Nearly all the ad technology innovation in the last decade has focused on what to do with that other half in order to raise the median price of that revenue from nearly zero to a bit more than zero."

Want to know which papers and magazines scored the best or worst in our on-going work?  Give me a shout at Happy to discuss the details with you.  

And stay tuned - more results coming. 
Add a comment...

Statistical Content Modeling

At ADmantX we have been futzing about with statistical content modeling the last 2 months.  And we think we have worked out some interesting ways of understanding a priori the likely ad matched to content.  In fact we can now do this in an A/B test environment.  

For the A test we model the keywords in content and from a representative set of ads running in a domain.  A virtual ad server makes the ad to content match and we record the placement.  Then we run the model again plugging in the keyword variables to test the accuracy of the model.

For the B test we repeat as above but this time we model the content and ad taxonomy using semantics rather than keywords.  This turns out to be a much deeper representation of both.  You would expect the accuracy and the reach of the B experiment to be much better since it can segment better as well as match across a wider range of content.  

In fact this is exactly what happened.  The reach of the ads using a semantic instead of keyword approach improved by a factor of 8-14 times.  And the accuracy improved by a factor of 2-3 times.  

Now that is an ROI you can take to the bank.  
Add a comment...

Post has attachment
ADmantX will be the only contextual data provider showcased at The Times Center during the Advertising Week Experience (AWE) this coming October 1-5, 2012 in New York City.  We are proud to have been selected for this prestigious week long event that includes the worlds most important ad technology companies.  

Be sure to come see us during AWE this fall at The Times Center 

242 West 41st Street
New York, NY 10036
Add a comment...

Post has attachment
Here is the message we have been slogging for over a year now and one of the primary reasons ADmantX exists.  Forking the emotional resonance out of content is no easy trick but once you do we contend you get 3 great benefits.  

Foremost among these is the  "last mile" of connectivity to the consumer.  What I mean here is that what the consumer is reading at ad placement time is final piece of the puzzle in understanding their frame of mind.  An hour ago, a week ago or a month ago based on historical behavioral tells us little about what the consumer is thinking and feeling right now.  But the content does.  So use it.  

Second is the inclusion of the content profile as a barometer to the stack of behavioral data.  Customers ask us all the time to go back through their log files and add new data based on the pages consumers have visited.  It sharpens their understanding of individuals as well as more firmly defines which groups individuals belong to.  

Last is the emotional trending among all consumers and content.  This is really the big data play but from a creativity or emotional side.  Why shout out happy shiny people ads when the country as a whole is a foul mood?  Again content analyzed at large scale gives such moving indications and can be used to tailor the creative side of ad creation.  

Bravo Paul for joining your voice to ours in linking creativity to the ad tech world.  
Add a comment...

Post has attachment
We have always liked Pubmatic.  They "take sides" with the publisher as their tag line tells you.  Working to preserve the value of the content creators is a smart move ... "rescuing newspapers" and other things most of us can get behind.  

Keep going Pubmatic.  You are the bigger version of us at ADmantX, working on getting the most from the worlds content for publishers, readers and advertisers alike.  
Add a comment...

Post has attachment
I am wondering today how the best and brightest pontificators in the ad world can miss the most obvious of connections over the last 2 days.  Mystifying.  

Yesterday we learned Microsoft would make Do-Not-Track an automatic option in their new browser.  See just one of many pieces on this event here.  And today during the DoubleClick live event we learned Google wanted to take over the rest of the ad world they don't already control.  That news is here. Actually the Google piece may not be real news to most.  Google has not really hid their ambition to suck up the $40 billion a year in spending they don't already get their hands on.  

Its competition ... plain and simple.  Google has the size and financial resources to make a real dent in that $40 billion.  Microsoft simply cannot sit idle and let them do it.  As a major browser company that can shut off the data spigot by making DNT the default option that ships with IE 10 and trust that most consumers won't turn it back on.  

But wouldn't Microsoft be shooting themselves in the foot as well?  Maybe - or maybe Microsoft has a second surprise up their sleeve.  Can you say operating system?  
Add a comment...

Post has attachment
One step closer to our prediction that browser makers will make Do-Not-Track easier and more universal.  

But still a better solution would be to prompt users with a choice the first time they open the browser.  Then the will of the consumer would be clear.  

The argument here that a default DNT choice in IE10 is equally valid for the DNT choice not selected by default as with most browsers now.  You cannot argue it both ways.  

And while we are at it here is a sub-prediction.  Should the day all browsers present large clear Yes or No buttons on behaviorally targeted ads during first run you can beat there will be a huge fight over how to word the choices. 
Add a comment...

Post has attachment
Very nice summary for those who need to get up to speed on the complex ecosystem of advertising technology e.g. the unwashed masses.  But also very useful for those of us in the system with our head down doing our thing.  Afterall "the kluge" is ever changing.  Nice job.  

Please include ADmantX next time in your list as a contextual targeter.  Thanks. 
Add a comment...

Post has attachment
Join us for Can Brand Advertisers Use “Big Data” AND Stay Creative?

Reaching your target audience online is more difficult than ever before. The main issue limiting the display market is the disconnect between advertisement placement and the relevance of the content. This is due to the reliance on keyword frequency without considering the meaning of content and the feelings and emotions expressed in content that could be transmitted to readers. Ads may appear on pages that have little significance to the assigned page or may produce counterproductive effects and traditional ad targeting continue to fall short, then.

Emotion, motivation and behavior are the means to human connection – what connects them to each other, concepts and actions. These can be found in content without cookies or other privacy invasion means. And these analytical components of ad targeting are the key for brand advertisers to stay creative.

Semantic advertising is bringing creativity back to advertising with the introduction of semantic targeting that incorporates emotions, motivations, behaviors and buying intentions.

You will learn more about;

- the evolution of online advertising, contextual advertising and the role of semantic technology in online ads placement.

- the benefits of semantic targeting.

- what information you can extract and use for optimizing ads.

- the use of sentiment analysis.
Add a comment...
Wait while more posts are being loaded