Posts Tagged ‘data analytics’


There are so many data providers and they’re examined and chosen by files, size of company providing, white papers not segments and results this is a big problem and just one of many.

When I started in the data business we made the mistake of using licensed data from all 3 credit bureaus, 4 county record providers, at least 4 demo files, 5 transactional files and even more…What we learned after much testing is all the providers had value but in certain elements not the entire files. We paid for so much redundant data to find the right elements to build our data products.

These problems created an entire business unit, we were very lucky to be serving the financial markets which paid top dollar for modeled data products and allowed us to still turn profits after paying all the license fees but after years of building our own products from these files we found in each file components of what we needed and cut out the rest of the files.

The process was intense and needed a lot staff heavily educated and experienced in modeling and data science to build products to test and then cut the necessary elements and coverage. The amount of variables were insane from geographical coverage, criteria validity, hygiene, data freshness, audience reach and match rates. We also needed to include the sales team to ensure we were building something based upon real client feedback and demand.

The next step was my partner and I renegotiating terms with our data providers from full licensing agreements to partial segment agreements…only paying fees for the data segments needed to create our unique products. We had to prove there was more financial upside to on demand and segment licenses versus full file installs. Not an easy sell considering the ease and profit of dropping off a full file every week, month or quarter and having us do all the hygiene and dedupping. We learned omitting data was just as important if not more than allowing data.

We negotiated based upon update cycles, batches of elements, higher prices for more utilized data sets and selects, and even longer term agreements if we had to, we knew this was risky but we saw the reward and huge unique selling proposition for our company. Some vendors laughed, some played ball, some took advantage (or so they thought) with higher pricing…our key to the pitch was our training of our sales reps to sell data the same way we needed to buy it and would build it was going to make us all more money. We were going to productize data in a way never done before.


There is so much more data now traditional, cookies, ip, set top box, survey, network and social but I see no difference in how it is needed to be purchased and sold. However the difference I see is now it’s my job to get third party companies to follow a business and sales model we created versus my own employees. I need their trust that it will work and their clients will see better performance.

I look at the adtech and martech space and still see these remnant buys with some audience data packaged but not enough lift in pricing and audiences that are still way too large to serve too…so too much data and not enough emphasis on a sales team that understands and can pitch the value in data elements being muti-compiled to create smaller audiences to create higher returns for the providers, the servers and the end user all simultaneously and this is possible…it’s being done but not enough.


The data providers are not offering the two key services needed to make this harmonious success theory a reality…training and support. Constant training and support.

Here is an example of the inability for me to buy data from a third party:

I logged into a native ads platform I know pretty well to launch a B2B campaign for a client of mine, simple steps went smooth, loaded the advertorial, pictures, URL etc. Below was the select audience data immediately I selected a source I am very familiar with I have worked with their data for years. So I began with 3 selects and pressed NEXT, the spinner started and just kept spinning until it crashed…OK mistakes happen so I started it all again and it crashed again. I ended my attempts to run the campaign the third party and the source lost utilization from me.

The data was not installed properly, the match process was not reaching out and grabbing right and the MVC was not stable so no sales and no campaign for my client…everyone loses.

The best examples to this training model I can give are Oracle and Salesforce, two companies that created products that created huge value for them and their clients by having powerful consulting businesses constantly providing training on use of the products and the data the products provided.

So now these two giants are buying the data companies and slowly attempting integration into their existing software products while also creating separate data sales business. The wall they are hitting is lack of utilization and ability to sell premium data products. The data is being utilized by companies trying to increase performance but not the price for that performance. This makes 0 sense to me, quite simply you buy a Porsche Carrera at one price and a Turbo at a much higher price…the turbo performs better and offers more…and it’s evident.

Building better performing data products and audiences if done right should be evident right down to the bottom line of both the provider and the end user, which means both parties should increase price for value.

So how do we change what needs to be changed to make this happen? Same way Oracle and Salesforce became multi billion dollar companies, provide the product and train the end user on how to best utilize it and sell it. Sales training by the vendor an old software model that needs to be packaged and delivered properly for data.

I remember years ago at Red Clay when the reverse mortgage business started booming and we were crushing direct mail response rates for a solid year, then those response rates started declining fast. We needed to change the audience. So we took in a ton of campaign data the positive data and the new negative data and looked at the changes…the ah ha moment came when we needed consumers with more debt management issues. So we had to use one of our vendors that had more late payment coverage. We immediately switched the late payment criteria from one bureau to another and instead of 60 day lates we went with 90 to 120 day lates…BOOM response rates went back up in tests.

Now we had built a new and improved data product for reverse mortgage that would double our competition’s response rates, next was a series of sales meetings that included all staff involved with product creation to create new pitches that documented how we built the file and the new and higher price we had to demand for it because with these response rates clients would double their profits.

This example is so agnostic to any data product/service, any audience, any use of data from traditional to digital to research and planning utilization. But it takes this holistic approach of all the teams working together, collaborating and being intimately involved from problem to solution to really embrace, trust, get excited about and sell the product. This requires training!





I am data. You are data. We are all data…and that’s why we have to zone in on our processes of data analytics and science as people because that is what the data is, people.

I am not alone here. I have had conversations with many of my peers and competitors who are all fully aware of how we must take an emotional approach to our work. This is because there are emotions and at times, very important life events that we are affecting through enablement and change, hopefully with the best of intentions.

What college should I attend and why?
Predicting intent to move, the changing of home caretakers, purchasing expensive items that will have a big impact budgets, choosing your next job – all these decisions affect so much more than a click or a day of someone’s life. We are mapping the flow of this traffic, its direction and the related content. Our research data, when brought to action, becomes the behavioral data about the people who make these types of decisions.

I feel the pressure of the people, not the CPM! 

Marketing, advertising and research through audience science, development, management and optimization, are all very important. However, without an emotional understanding of the people and the goals, we can easily fail.

Even understanding and creating the right content might not be enough to accomplish goals focused on the people – not just a CPM or CPC. The IP models get loaded into the DSPs for programmatic media buys and we can often forget they are actually people. We can no longer run those buys the same way we have watched display advertising become a commoditized market of price and reach, not quality and relevancy. We need to bring favorable and tolerable marketing for the people!

But here is why advertising has the opportunity to really harness the power of the people – by really listening to the that data, analyzing it, finding stages in the buyer’s journey and making it speak to us. However, if you don’t engage with the proper content that says “I hear what you want and we are giving it to you through incentives that increase our life time value through a data/people relationship”, then we are not committed to a greater outcome and relationship.

We can bring the horse to the water from audience analytics but will they have a drink?

We need to emphasize:

  • Creative solutions
  • Relevant content
  • Offers and meaningful CTAs
  • Lead and prospect nurturing through real-world CRM solutions

The ‘Relationship’

It must be a fair relationship between your data partners, ad agencies and company processes. This how we get to a win-win situation, for both the business side and the consumer satisfaction side. This is also why it never matters who the smartest guy in the room is it’s the contribution of each individual – their skills and experience independently based upon the goals we are all there together to achieve. The lions need to form Voltron.

As for us data, audience, people profiling folks – we can help by doing our part. However, it’s rare to be able to have an audience scientist who is skilled in UI, UX, design and content. This is an illusive mash up and usually is only found in the old school direct marketers that have evolved with data for over 15 years plus of experience. This is because back then, it had to be done this way. The creative in DM/DR was responsible for processing the data and then communicating the meaning of these analytics back to the creative – much more than A/B testing.

Insights are only relevant in advertising and marketing if action can be tied to them. Data can be extrapolated from most of the insights, especially from research data and not old survey panels. Keep in mind, I don’t just mean for media planning and buying. This applies to making actionable decisions based on locations, banners, landing page optimization, customer programs, sales times, seasonality, AdWords, competitive analysis, consumer experience and much more.

Intent, from cookies to IP addresses, started and has been evolving since about 2012.

Nowadays, the cookie is basically irrelevant in intent modeling and audience development for so many reasons, mobile being of course the big one. There are no cookies in this terrain and all around the cookie…well it’s just not clean data in! So now, we need to uncover real predictive intent and this comes from research analytics and modeling that can be tied to indexing page consumption at the UA/Device level. Actions can then be taken from all aspects of communications and engagement, creating many important triggers.

Data is becoming more and more useful for practices such as Venture/Private Equity, PR, Product Development, Market Analysis (sizing, opportunity, competition, etc.). I cannot stress enough how we can no longer just load IP addresses into a DSP and start programmatic buying with old methods. While this is an important use, it needs proper recency and frequency models attached to it for optimal performance, in brand as well as direct. Many are busy in their labs as we feel, with so much data these days, there is limitless work to be done.

The bottom line

While it’s really fun to see what comes out of an audience profile in the early part of the build and then, how many segmentation and utilization doors get opened, I often find that it’s very easy to get lost in the data.

So how do we maintain a clear focus?

First off, if we are tasked with a project that has a defined end result, this helps a great deal. Additionally, being focused on improving campaigns, strategies and ideas all of which are based on the who, the what, the where and the when, provides you with a clear road map to follow.

With so much research and work being done to avoid digital ads all together for consumers, it’s clear we’re not performing optimally.

As Marketers, we have the power to do better! And therefore…we must accept the responsibility.