Improving the Lives of Patients with DeepIntent

For Chris Paquette, Founder and CEO of DeepIntent, the creation of his company is the combination of his past experiences in academic research as an engineering student and his work. “Prior to founding DeepIntent, I worked in ad-tech and later as a data scientist at Memorial Sloan Kettering, working on machine learning-based solutions,” he explained. His empathy with the patients there led to a revelation. “I often spoke to patients about how difficult it was for them to learn about disease diagnoses and treatments. I connected deeply with their perspective because when I was 12, my father was misdiagnosed with the wrong type of cancer. His gut told him something was off so he got a second and third opinion, which ultimately saved his life. Realizing how powerful machine learning could be in terms of getting the right information to people,” he noted, “and this was the genesis of DeepIntent, where everything is driven by a core belief that advertising technology can measurably improve the lives of patients.”

His company leverages a range of measurement tools from unique datasets, machine learning and advanced algorithms. All of these are making it possible for advertisers to interact with doctors and patients in a privacy compliant manner according to Jen Werther, Chief Strategy Officer for DeepIntent. For her, DeepIntent is, “A healthcare marketing platform. We serve ads on behalf of pharmaceutical companies to either physicians or to the patients,” with a marketing platform that is built on three pillars. “The first is Plan, where we have unique data sets to create custom audiences for physicians. Then Activate, Measure and Optimize where we use the unique data to plan different audiences for our clients.”

DeepIntent boasts 1.6 million NPIs where, “Every physician in the US has an NPI number that's unique to that identity and we've built out an identity graph of NPI to digital identity cookies and what we do on the patient side is we have unique data to create patient modeled audiences understanding whether this user look to have a high propensity of being a Type two diabetic (for example),” she stated.

Data is used prudently and strategically. “We source a lot of different, unique data sets to plan different audiences. This goes directly into our DSP which has been built in-house. Everything is proprietary to us,” Werther explained and added, “So we serve ads to either physicians or patients across the Web, display, mobile, audio, CTV, native, all of the digital channels, and then what we have is measured and optimized with the ability to optimize in real time towards scripts. Did I reach a patient who ultimately got a script and did I reach a physician who wrote a script whose patient actually got a script?”

Other data sets that hold great information are insurance claims and eligibility data. “We have access to medical and pharmacy claims which we use in a couple of different ways,” Werther shared. “So if I want to understand those diagnosed with Alzheimer's in the last six to 12 months, I could create a custom audience based on an ICD10 code, based on the diagnosis code, to look at medical claims and come up with the list of physicians that have the most patients with that diagnosis. Or I could use pharmacy claims to look at physicians diagnosing either my brand or the competitive set. You could tier them into high medium low. That's what pharmaceutical companies do. Medical and pharmacy claims allow us to create custom audiences using diagnosis codes and procedure codes.” All of these can also be targeted to a specific geo location, state and DMA.

For Parquette, accurate data matched with privacy compliance is paramount. “Getting the right information to people requires a degree of personalization. This is inherently challenging in the healthcare space given that regulations like HIPAA prohibit one-to-one targeting based on health data. Because all of DeepIntent’s technology is built purposefully for healthcare marketers, we build new solutions with those considerations in mind, enabling healthcare and life science companies to reach the right patient audiences in a privacy-safe way,” he stated.

Werther noted that on the patient side, “We model audiences using medical and pharmacy claims. For example, I look at patients who've been diagnosed with Type Two Diabetes and use that to create a predictive model. We’re not targeting patients one-to-one. Maybe I want to find caregivers so we have third party audiences that look at people who have engaged with Alzheimer's content. If I'm an Alzheimer's patient, I'm probably not going to understand if I'm seeing an ad on the web. I want to reach the caregiver who's doing research on Alzheimer's.”

For agencies, DeepIntent offers tools to facilitate measurement and optimization quickly. “We set up business rules to say, I want to understand this set of physicians before they see someone with a rare disease. With claims data, it already happened. Eligibility data looks at the future. I'm identifying physician who are about to see a clinically relevant patient and if you think about rare disease those patients may see that specialist once or twice a year, so it's really important that you get that message in front of the physician prior to that ad being shown,” she added.

In terms of locations, DeepIntent also has access to, “One to one mapping based on a personal email or work email. So we're taking all digital identifiers of an NPI to be able to come up with an identity graph. If I'm a doctor, I could be reached via a display ad, if I'm on CNN or Yahoo. But I also could see a message if I'm on an endemic site like Webmd or Healthline or Health Grades,” and, she noted, “We don't serve inventory on electronic health records. EHR inventory is not part of our inventory today although it could be in the future.”

Optimization is a priority and a game changer because while, “Marketers typically optimize towards clicks and impressions. What we did is build an optimization engine,” that relies on machine learning and proprietary algorithms that enable DeepIntent, “To optimize in-platform versus doing third party measurement posts for that campaign,” she shared.

When it comes to the future, Werther understands the need to innovate. “Our future plans are to continue to innovate, to be the market leader in healthcare and find new data sources. That's really my passion and a lot of the work that I do at DeepIntent is finding unique data sets to help us target either physicians or patients,” she said.

“Technology is constantly evolving, which calls for tech companies to evolve in lockstep. Since I founded DeepIntent six years ago, the company has changed quite a bit, but our core mission has remained the same, which contributes to our strong retention. Clients have the confidence that as we continue to innovate, we’re still fundamentally driven by helping them help patients,” Paquette explained.

“To me, a perfect version of DeepIntent would consist of us doing exactly what we’re doing, just on a much larger scale. When we surveyed patients last year, we found that 50% are more likely to take a treatment recommendation for something they recognize from advertising. Getting the right information to people drives awareness and fosters more open conversations with their healthcare providers, which leads to a healthier population. Contributing to that is DeepIntent’s ultimate goal, always,” he concluded.

This article first appeared in www.MediaVillage.com

Artwork by Charlene Weisler

The Brave New World of True Accountability. An Interview with A+E Networks’ Roseann Montenes

Can we ever get away from the antiquated demographic guarantee?  According to Roseann Montenes, Vice President, Precision and Performance Advertising Sales, the answer is a resounding YES! And she should know. 

Montenes is one of the forces behind A+E Networks’ MVP (Multi-Viewing Precision) initiative, a system that moves marketers from legacy metrics to true KPI accountability.

Convincing the Industry

But there is something a little unnerving to advertisers about moving from the comfort of demographics to true accountability. “It is essentially like moving a small mountain,” she admitted. “You have clients that are so used to doing business one way since the dawn of time. It was almost like when we made the shift to C3 and had to get clients into that mindset,” she added. The shift began to occur after, “tons of education, tons of case studies that demonstrated to clients that they could take the demo guarantee out of it and focus on who their true audience is.”

The first step is to, ”prioritize all of the client’s KPIs to see what matters most. Is it targeting households that have anyone from adults 18-49 in it? Or, is it targeting someone who is in-market to purchase a new vehicle?” she explained. The answer is usually to move product. “At the end of the day, that’s what matters most. That is how we started to change the conversation.” And yet, even with this great effort, it took a full broadcast year cycle to get a client onboard, comfortable and then implement.

How MVP Works

MVP offers marketers true cross platform optimization. But how does it actually work? “What we do is take a strategic target audience and optimize through our Precision platform, which is our linear arm, find the audience there. We then do an optimization on the digital side through our Digital Precision offering and then we blend those two plans together,” she explained. “We allow the client to have full access into the allocation of impressions within the linear and digital arm and then it is figuring out the balance.” Notably each campaign is monitored by hand to carefully consider how each spot, whether linear or digital, contributes best to the KPI goal. Both the linear and the digital teams are in constant contact to make sure that the overall plan is on target.

Delivering the KPI

What is the most meaningful metric to offer in place of demographics? For A+E Networks it is viewability. “We just did our first guarantee against attention measurement in fourth quarter,” noted Montenes which was based on attention measurement through partner TVision. She added that MVP has the ability to ingest any data set. “Our management doesn’t limit us to what kind of datasets we can partner with. Any third party dataset, we can onboard an advertiser’s own first party data onto our platform and any other standard, generic, statements whether it is a Polk dataset or MRI. For those advertisers who don’t have their own strategic target in place, they can come to us and we can help them create one,” she explained.

To craft a plan, MVP enables a drill down, “to the sell title level, to the half hour level. This allows us to make those in-campaign optimizations based off of the pacing of whether it’s our attention measurement or a guarantee against a business outcome, or whether we are optimizing against reach or a particular strategic target,” she stated. This enables the network to maximize the value of all of their programs, regardless of demographic performance levels. “There are no two programs that are equal when you are talking about a customized strategic target,” she noted, as opposed to being able to compare programs based on demographics which are more standardized but less targeted.  

Impact of the Pandemic

With one major advertiser on board, the teams at A+E networks have been able to monitor a full broadcast year worth of MVP performance. “Now that we feel comfortable and know that it works, this is something we are going to bring to other clients as an opportunity to change the total outlook of currency and how to trade off of it,” she assured. Of course the landscape has changed since fourth quarter 2019 with the pandemic but a marketer could successfully pivot… and should. Montenes explained that, “For a really long time, we were all about foot traffic into store locations, dealerships and into family dining establishments. Once Covid hit, we felt this sense of ‘oh no’ now that foot traffic has stopped, because you are not allowed, how do you reimagine what a partnership looks like?”

After the initial sense of panic due to the pandemic dissipated, “it quickly fixed itself,” she stated. So instead of actually dining in a restaurant, consumers were downloading the app. “We switched the conversation and KPI to be reimagined and looked at differently. It went from foot traffic to app downloads and foot traffic to conversions.” Notably in some cases, the advertiser still wanted to deliver on foot traffic.

Going Forward

For Montenes, there is a new normal for media now. “Clients have gotten a peek under the hood in terms of what their business could look like. There is now efficiency in seeing the world that once was and measuring it in the world that now is. How can you go back to the way it was before now that we see what the new world actually looks like?” she asked and added that she is very optimistic about the future. “It’s exciting to see how much has changed since last March. It hasn’t changed for the worse. It’s changed for the better,” she concluded.  

This article first appeared in www.MediaVillage.com

 

 

 

NCS's Leslie Wood on Monitoring Campaigns Using AI Causality

High on the wish list of most marketers is improving the measurement of campaign effectiveness. Fortunately, Leslie Wood, NCS's chief research officer, has been on a mission to perfect it. She has accomplished this with a combination of machine learning methodologies that interface with a campaign in flight, thus enabling marketers to modify and improve their ad placements in real time. This initiative, called Sales Lift Metrics, sheds light on the causal impact of specific campaign tactics by identifying the key sales drivers that amplify incremental sales for CPG brands.

"We have created a Super Learner, which includes multiple different machine learning methodologies, like 'random forest' and 'gradient boost,'" Wood said. "And when we execute the Super Learner, it runs all of those different models and sees which ones are winners — often up to three or four of them — and combines them into an ensemble model." It is apparently flexible enough to apply to any type of CPG advertiser and is constantly refining its application through the use of AI.

How It All Began
The idea and the efforts behind this initiative started about five years ago and, over that span of time, the methodology has evolved through testing and innovation. "When we first started, it was just television because you can't have a control if you have such high reach. After all, who is not exposed? People who were not exposed to the advertising look very different than TV viewers. They are very different kinds of people," she explained.

Wood discovered that this approach took the pressure off the researcher to constantly oversee the model. "Here was this machine learning method that allowed us to put this in places where data was siloed. We could set it up in 'free rooms,' places where there was no researcher," she said, adding that this allowed a door to open where Google, Facebook, and Apple could manage their proprietary datasets without a human interacting or overseeing it. "It is meant for computers to talk to computers."

In-Flight Optimization
Five years ago, the hardware just wasn't there to be able to apply in-flight optimization. But now, this can be executed on what Wood calls, "sales-lift metrics," which provides the ability to "use these machine learnings to quickly, in real time, deliver key performance indicators to tell you what about your campaign is working." It is now possible to compare multiple creative options, consumer targets, or sets of tactics for sales effectiveness and incremental causality.
"We are looking at the incremental value of each of them," she said, "and compare them" in a weekly cumulative sales report. An advertiser can make changes to the campaign as it rolls out and see the impact.

What Sets the Super Learner Apart?
"There is always a full set of statistics" within the Super Learner, Wood noted. This is unusual because "AI almost never has a full suite of statistics. Statistics were developed by professors on small data sets. For big data sets, we develop machine learning and AI methods and those were built by computer programmers. So, we don't usually have confidence intervals, significance tests, or standard error — those measures that you need to say that this is a good model. We built Super Learner, which does all of these steps."

In this way, as new machine learning methods are developed (and, according to Wood, they are being developed all the time), they are added to the system to see if there are new ways to improve campaign efficiency. "It allows us to continually improve without reprogramming or recoding," she said, adding that certain models never worked and others that worked too hard and delivered unreasonable answers. "We removed those."

Among the successful campaign indicators is incremental sales per household, which is a key performance metric. It also reports impressions delivery by target and by creative.

Going Forward
Since the Super Learner has the ability to integrate faster systems and new models, it advances every day. Thousands of new schedules have been run and rerun "to see what works and what doesn't work, what's the minimum threshold, what's feasible, what are indicators inside the statistics that let us know that it's not working."

Wood is committed to keeping up with "this fast-moving world" of data and machine learning with constant monitoring and asking why results change. In the fall, they will be adding more data, more brands, and a price/promotion feature.

And, according to Wood, this entire initiative has been a model for others. "We use the h2o machine learning language; they will tell you that we were the first to create a Super Learner and they followed suit and built one themselves," she said. "We have pushed the envelope on machine learning."

 This article first appeared in www.MediaVillage.com

Advancing into Linear Television Optimization. An Interview with Stacy Daft of Amobee

Linear television is currently grappling with how to optimize its offerings in a media world that is increasingly leveraging granular data for targeting purposes. 

The recent announcement of Univision and its selection of ad platform, Amobee, will, according to the press release, enable advertisers to plan and transact against their custom strategic target audiences or proprietary first party data creating a foundation across linear, digital and social channels.

“Univision is the largest broadcaster serving Hispanic America, and has been the number one Spanish-language network in primetime for 27 consecutive years,” explained Stacy Daft, GM  Commercial Business Development, Amobee. “They command a 60 percent share of the Spanish-language primetime 18-49 audience, reaching an estimated 106MM unduplicated consumers a month, with 90 percent of their audience watching in real-time,” she added.

Daft offered the following insights on the partnership and the platform’s attributes:

Charlene Weisler: What do you mean a by linear television optimization platform?

Stacy Daft: The Univision / Amobee partnership advances data and insights across linear TV, bringing the better understanding of audiences, targeting, data and insights we’ve come to expect in digital. Univision can now provide a data-optimized linear television offering to advertisers by applying the same principles of granular audience targeting and campaign measurement as digital, allowing advertisers to better understand audience reach across multiple networks and plan and transact against their custom strategic target audiences or proprietary first-party data.

Weisler: What metrics are used to measure linear planning?

Daft: The primary metrics for advanced linear planning are strategic target index, reach, frequency, impressions and CPM.

Weisler: Is it programmatic?

Daft: It’s not programmatic; Amobee is not operating as a DSP connecting to Univision’s demand. Amobee is acting as Univision’s sell side enterprise platform, optimizing for their direct sales business. With Amobee’s technology and software, Univision will create data-driven plans to help advertisers reach strategic targets across their portfolio of networks.

Weisler: How is it innovative?

Daft: Univision is an extremely innovative broadcaster and this work to make linear television more data-enabled is the type of partnership that helps move the entire television industry forward. Bringing data to the top of the marketing funnel provides better targeting and addressability to television investments, with overall television advertising spending slated to surpass $200 billion by 2020.

Advanced linear television optimization brings more precision to the buying process, allowing targeting beyond age and gender and allowing greater automation in the buying and selling workflow. By layering first-party and third-party data over television viewing data, media buyers and sellers can understand more about reaching their best consumers, most efficiently across Univision’s linear properties.

The relationship with Amobee enhances Univision’s ability to create end-to-end solutions for clients who are seeking advanced data solutions and precision targeting against the valuable Hispanic-American consumer segment. This partnership signals our commitment to collaborating with major broadcasters to deliver best-in-class digital and linear TV capabilities that help bridge the gap between digital and TV and are specifically designed to enhance the TV advertising proposition for advertisers and agencies.

This article first appeared in www.Mediapost.com