#datavalue

AI Starts Working Better When Data Engineering Companies Get the Foundation Right

Strong data systems are the foundation of modern decision-making.

Data engineering companies focus on building reliable pipelines and ensuring data consistency across systems.

Without this layer, even advanced tools struggle to deliver value.

For instance, take the category of pickup trucks. Most brands think their audiences are the male, cowboy types. In reality, 40% of pickup truck owners are women & only 1% work in the agricultural sector. Brands like Ford & Chevy are making guesses over and over again & usually get it wrong. Or take Walmart, one of the largest retailers in America, and they like many retailers hone in on Moms as their main target because moms have all the buying power for the home right? In reality, 50% of all Walmart shoppers aren't even married and most are males. In the world of big data, we no longer have to guess. We have the tools we need to be right, every time, but we haven't built the masterpiece to mining this data: the database of affinity. What is the database of affinity? People's preference for - or desire to connect with - other people, products or things.

Liking & voting for things - 83B likes on Facebook per month, 1.5B instagram likes per month

Talking about things - 12B tweets per month, 39M Wordpress blog posts per month

Reviewing things - 1M Yelp reviews per month

Spending time with things - 4B hours of YouTube video watched per month, 110M Foursquare check-ins per month

Sharing things - 77B Facebook shares per month, email is the #1 way people share content

The power of this database paired with existing databases of intentions is unfathomable when you think about it.

The real value of social networks is in the data consumers create, not the ads. This data can power advertising everywhere. #datavalue #sxsw

March 10, 2013

What is a database of intentions? A catalog of people's needs and desires collected by observing their search behaviors.

Sounds familiar right? That's Google.

So you may be thinking, well couldn't Facebook be the database of affinity? Well Google made 50B in revenue last year while Facebook only made 5B, and they have more users than Google. Is Google just better than Facebook or is the kind of data we offer just different?

Let's look at the differences in the data:

Intent data is observed by search engines

Affinity data is observed by social sites

Intent data is expressed by exploring

Affinity data is expressed by engaging after the purchase

Intent data is rational. Affinity is emotional.

Intent data last days or months.

Affinity data is evergreen, lasting years

What does this mean for us?

Intent databases help direct marketers.

Affinity databases help brand advertising. So who can build this affinity database?

Three things are needed to build a usable database of affinity: 1.) Data from across the social world 2.) Analysis tools that bring meaning to that data 3.) Ad formats that create brand impact

The database of affinity can help brand advertisers reach the right people at the right time. #datavalue #affinity

March 10, 2013

Why Forrester thinks Facebook can't do this? 1.) It relies on one single source of data

Facebook's own affinity data has key weakness. It isn't obtaining affinity data from other sources and don't know where to go/start.

2.) It struggles to bring meaning to data

It has little experience figuring out what content users like.

It doesn't have a good mechanism for validating its analysis to prove what's interesting to people. The like button isn't enough.

3.) Its ad units can't create brand impact

It offers no video ad units. Its new units are still inferior to standard banners.

Facebook affinity is a measure of aspiration more than a representation of actual customer behavior #sxsw #datavalue

March 10, 2013

Why Forrester thinks Google can build this:? 1.) It collects affinity data from many sources.

It already collects much more affinity behavior across platforms than you think. YouTube, Google+, Gmail, maps, and more. And if they don't have it they knows where to go for more. For instance, every check in on Foursquare is powered by google maps and every review on Yelp.

2.) It can bring meaning to that data.

It has lots of experience figuring out what content users like (page rank/search algorithm).

It has better mechanisms for validating its analysis. It knows if the ads they sell get a user to search for that brand.

3.) It offers the best brand ad units online.

Impactful video ads through YouTube and their TruView model that allows brands to only pay for what users watch

How could this change Marketing as we see it today?

Forrester predicts, in the next 2 years, we'll start to use more data more wisely. Data sets will start to expand and we will move past single-variable affinity targeting so you can start to understand what a combination of affinities really means. By 2015, we will start seeing complete social data sets and impactful ad units. Within 5 years, affinity data will start to take over TV advertising as we know it.

Affinity Databases will roll out in 4 Key Stages: 1.) Isolated data sets with little data analysis, self-serve buying tools, & low impact online ads 2.) Smart affinity with broader data sets, multivariate data analysis, ad exchanges, & low impact online ads 3.) Big affinity with comprehensive data, multivariate data analysis, ad exchanges, & impactful online ads 4.) Ubiquitous affinity with comprehensive data, multivariate data,  algorithmic buying, & offline ads

My Thoughts?

Building a database that links affinity data with intention data is crucial to the survival of our industry. Sure MRI, Forrester and the likes will always be valuable data sources to guide our marketing plans. But a true data base that can accurately depict true affinity and intent for a brand is the silver bullet we've all been waiting for.

Facebook's Open Graph search is one attempt and nod in this direction. They understand the power of this data, but unfortunately they are only one source (albeit a huge one) of data. To paint a truly accurate picture of applicable targets for our brands, the data has to be pulled from multiple platforms. It cannot just be MRI, or Facebook, or sentiment on Twitter or even transactional data. It has to be everything.

Simple right? ;)