Industry Insights
Into the Looking Glass of the Data, Insights & Innovation Team at Mediavest

Lenny Chase
previously Data, Insights & Innovation @ Mediavest
now VP, Digital Marketing @ Halo Top Creamery

 


 

Please tell us a little more about your previous role managing Data, Insights, & Innovation at Mediavest and anything you are working on now.

Sure! Let’s start with some context: The role was created in response to not only the need, but also the speed at which advertising technology developed. More importantly, the question remains—how do we leverage and implement these developments? Our team’s goal was to help answer this question. I think the best way to explain my role is to compare it with product management in the tech industry. They are very similar in the sense that I had to put on many hats and work with many teams both internally and externally. The product in question was a first-party “DMP” of sorts and attribution architecture to tie digital impressions to in-store sales. It was called WMX. (http://www.adweek.com/news/technology/473-billion-retailer-wants-be-next-ad-tech-star-161471)

Internally, my team evangelized this tool among the planning/buying teams and demonstrated how it leads to smarter and more efficient media. This covers everything from creating/utilizing audience targeting, to analytics and benchmarking how publishers perform in regards to real in-store conversions. Externally, we worked with Walmart Labs in the development of this tool to communicate market requirements, user feedback, and most importantly, integrate with the advertising ecosystem. We would have daily meetings with publishers and ad networks to discuss how we can get our systems talking the same language. It was an awesome experience having had the opportunity to wear these different hats in a fast-paced industry like ours. I have since left to independently consult, giving me time to pursue a few projects that I’ve wanted to start. One of them is quick, cheap at-home testing for common illnesses.

 

Cross-channel attribution has yet to be fully solved in digital advertising, with marketers having full visibility into a users’ journey and therefore being able to assign accurate credit to each advertising touchpoint. What are your thoughts on view-through attribution for campaigns, and are there any sophisticated means of attribution or assigning credit that you’ve found worked well for clients?

Arguably, every impression has a marginal effect, so it is nice to collect this data. I do look at view-through attribution on campaigns, but only to see if there is an interesting story to tell. I can’t recall a time where I’ve used this data to make a business decision though. Perhaps this is a consequence of working in an unsolved area as you have said. I’ve used the time decay model in the past with clients because it made sense to. Unfortunately, there isn’t one attribution model that works the best.  According to Google, the time decay model makes sense “if you run one-day or two-day promotion campaigns, as you may wish to give more credit to interactions during the days of the promotion. In this case, interactions that occurred one week before have only a small value as compared to touchpoints near the conversion.” To generalize, every advertiser should pick an attribution model that works for their business model.

 

It’s been stated before that the half-life of a cookie is 3 days, with 1/3 of all cookies being deleted in less than an hour. If this is truly the case, how accurate is 3rd party data, and segments that can be bought from data partners?

That would be a nice answer to have and I wish I knew for sure! The best I can do on my end is to do due diligence and research. I often ask these partners for their collection methodology and a meeting with their engineering or data science team. It helps me vet where the data is coming from and understand any potential caveats.

 

Programmatic has been a hot topic over the last few years, with trade publications like eMarketer stating that it will be the main mode of buying in the near future. What is your definition of programmatic, and do you think this will be the case?

I think this buzzword is way overused and often times used incorrectly. Programmatic is a concept and architecture where a system, given a set of defined constraints by the advertiser, will algorithmically determine either to show an ad or not. These constraints can include, but are not limited to, specific sites, basic demographics, behavioral characteristics, economic characteristics, etc.

There is a common misuse of the term “programmatic” to mean “real-time bidding (RTB)”. RTB simply adheres to the concept of programmatic. As such, I would even say that nearly all digital advertising is programmatic. Here’s a basic example: If you are choosing to show ads in the United States only, your media is programmatic.

The cynic in me says let’s deprecate this term.

 

Do you think that endemic or “content as a proxy” publisher direct buying will ever go away?

No, I don’t think so. As more advertisers shy away from share-of-voice targeting toward audience and people targeting, it makes sense to be in the context of where your audience will be naturally. This would further be supported by an increase in demand for native ad units.

 

How do you think this whole ad blocking issue will play out, especially for programmatically purchased media?

On the technical side, I see ad blocking here to stay, whether we like it or not. It’s our job to adapt and address the underlying issues. Why do people use ad blockers in the first place? Without going in-depth, ads can have user experience, privacy, and security implications. Digital media would be at a disadvantage (less targeted) in the short-term. Ad blockers would be blocking data variables that would otherwise be collected for use in the programmatic decision-making process. In the near future, I see a rise in native placements and the use of people-based identification (using first-party data) instead of cookie or device identification.

Ironically, I use an ad blocker for my daily browsing needs. My reasoning is as follows: Security-wise, there is a lot of garbage on “the internets” and malware rates have increased in the past years. Effectiveness-wise, I question whether or not the people who would otherwise use ad blockers, would respond to digital advertising in the first place. That would be a great study to conduct!  User experience-wise, it is unbelievable how many websites there are that have more advertising placements than content on a page. Before someone mentions “Well, said advertising is paying for the free content.” True, which is why I do disable ad block on sites that I frequent, but that brings up the effectiveness point. Although I’ve disabled ad block to support a site, would the cost of my impression paid by an advertiser, be better utilized on someone else who may be more likely to convert?

 

Are there any innovations happening now in digital advertising or data that excite you the most?

All the new digital mediums being created or transformed! VR is increasingly becoming accessible to the consumer and with that comes new advertising opportunities—or experiences. I would imagine others including myself wouldn’t be too content on standard banner ads appearing in my VR experience. New experiences call for new ways to advertise.

I’m also keeping an eye on technology that measures traditional media sources such as linear TV and subsequently is able to deterministically attribute conversions. This doesn’t scream cool, but nonetheless, a solution here would have a huge impact on advertising investment.

 

What are some ways marketers can leverage data, and tracking to attribute campaigns to actual in-store purchases?

This is a great and very important question that we’ve tackled in the past years. Currently, common methods such as probabilistic modeling using location data or survey data exist. Having implemented deterministic methods, I know we can do better than probabilistic. The minimum and hardest requirement to fulfill, is a link between the offline identity used in a purchase (e.g. credit card, loyalty card, etc.) and the online identity (e.g. account, browser, device, etc.). Once this datastore exists, this enables a full closed-loop attribution solution and enables targeting across the digital ecosystem. How cool is that? To reiterate, this isn’t a theory. We can do this right now. Not only did the advertiser I worked with architect such a system, but also we have implemented it in production and is now being used to track every digital ad impression.

 

When it comes to self-service advertising, will there ever be a point where there could be “One Platform to Rule Them All”, with access to most, if not all (including walled garden) inventory? Do you think any of the large platforms have this in their future roadmap?

While this sounds ideal, I think users would have to sacrifice the ability to fine-tune campaigns. If my theory, that native unit supply and demand will increase, holds true, then targeting options, delivery options, etc. will be unique to each publisher. Let’s take paid search for example: at one point, we could use a search management platform to manage campaigns on Google, Bing, and Yahoo simultaneously. However, as betas and new features get released, the standardization fails and we start reverting back to working in native UIs to take advantage of publisher specific features. I wouldn’t be surprised if large platforms have this in their future roadmap. The execution and user demand will be a discussion point for sure. With every self-service portal, there needs to be an operator or operators, which comes at a cost.

I thought of one model that may get us as close to that point as possible. An application similar to IntelliJ IDEA. If you’re not a developer or haven’t used this IDE to code, the application has a bunch of plugins that the user can enable. Said plugins allow the user to work with different programming languages or features, but the UI remains mostly the same. Instead of being languages, if these plugins were different publishers and inventory sources, and instead of an IDE to code, the application was designed to create and manage ad campaigns, that could be this “One Platform”. If you’re interested in building or actively developing an application like this, I’d love to connect! :)

 

Lastly, what are your thoughts on the future of the way media will be bought and sold, for both big brands/agencies as well as the mid-long tail of advertisers?

Long term without getting too complex, I imagine the way media will be bought and sold would mimic the way the financial markets operate. One goal should be to decrease asymmetric information between supply and demand. Similar to the financial markets, there should be an accessible order book which lists buy and sell orders for ads, and identifies the market participants. I also wouldn’t be surprised if a regulatory entity is created to address things such as high frequency trading on ad exchanges. As for the mid-long tail of advertisers, I think self-service is key for lowering the barriers to entry. In parallel, documentation and training material is increasingly necessary for all types of advertisers as technologies change and adapt.

Hero Image Source

Share