What Goodhart’s Law Can Teach You About Performance Data

Posted in Audience Series, Blog

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle topics to give a 360-degree view on this vast, ever-changing industry. The following article ran in AdExchanger’s Data-Driven Thinking series.

Roman Shraga
By Roman Shraga

Is there a metric you use to evaluate the effectiveness of something critical to your company’s success? What about a metric used by your company to evaluate you?

If so, it is essential that you understand what could go wrong in the evaluation of performance data. Your job depends on it!

Performance data is the information that is used to assess the success of something. It’s how you evaluate the effectiveness of an ad campaign, the throughput of an engineering organization, or the business attributable to a specific salesperson, for example. Because performance data is directly tied to the key goals of both individuals and organizations, it is a sensitive – and even contentious – topic. It is ripe for obfuscation and abuse.

Goodhart’s Law

A critical insight into how to deal with performance data comes from Goodhart’s Law: “When a measure becomes a target, it ceases to be a good measure.” In other words, when the measure being used by decision-makers to evaluate performance is the same as the target being optimized by those being measured, it is no longer a reliable measure of performance.

The most cited example of this law in effect is the case of nail factories in the Soviet Union. The goal of central planners was to measure performance of the factories, so factory operators were given targets around the number of nails produced. To meet and exceed the targets, factory operators produced millions of tiny, useless nails. When targets were switched to the total weight of nails produced, operators instead produced several enormous, heavy and useless nails.

The above example is absurd, but illustrates the point: When a measure of performance is the same as the target, it can be abused to the point of no longer being useful in measuring the desired outcome.

Advertising Implications

This happens all the time in the modern world. For example, when CTR is both a measure and a target, ad companies have a perverse incentive to optimize for clicks with absolutely no regard for whom is doing the clicking. An ad campaign for Ferrari with CTR of 15% sounds amazing — unless the majority of people who clicked the ads are teenagers looking at pictures of cool cars.

Similarly, when cases closed is both the measure of performance and target of customer service organizations, employees might choose to close cases without fully investigating and resolving them. When page views are both the measure and the target of news sites and blogs, editors have incentives to post shocking and controversial content to optimize for the target. In the long run, of course, this behavior degrades the quality of the site and the page views measure is no longer a useful indicator of the desired outcome of an engaged user base.

Mitigation Techniques

Examples of Goodhart’s Law can be found in every industry and every department of an organization. Fortunately, there are several approaches that can be taken to mitigate its harmful effects.

  1. The first approach is also the most difficult. By thinking deeply about what is being measured and what the constraints are, it is possible to formulate better measurements. A body of knowledge known as the theory of constraints can be used to guide your thought process as you try to come up with a better measure.

    For example, as an alternative to relying on cases closed as a measure of customer service, a company can learn from Zappos and strive to quantify and reward good experiences as reported by customers. Still, it must be said that there is debate about whether it is even possible to find a single measure that is immune to the effects of Goodhart’s Law.

  2. A second approach could be to create a “balanced scorecard” of several different measures instead of relying on one. With this strategy, you reduce the risk of a single measure being gamed by looking at multiple measures that evaluate performance from different angles. For example, CTR can be supplemented with a measure of traffic quality, such as bounce rate or conversion rate.

    When you add multiple measures to your overall performance evaluation, you not only reduce the opportunity for abuse, but you begin to get a more nuanced understanding of the inherent tradeoffs being made. This is similar to the dual metrics of precision and recall used in machine learning classification problems. Together they measure how often the machine gets the right answer and what proportion of the total right answers the machine is able to get.

  3. A third way to mitigate the effects of Goodhart’s Law is to simply use human discretion. This means poking and prodding a reported performance measure until you develop a true understanding about what it is actually indicating. You need to ask questions that ensure the measure relate to the ultimate goal.

    Additionally, think about whether it would be possible to get a perfect score on the measure, and if it would be possible, to do so without adding any value. This line of reasoning will allow you dissect a measure until you understand whether or not it is doing a good job of indicating performance.

In the end, a mix of all three approaches to mitigation is the most judicious thing to do. You should strive to create the best possible measures that look at performance from multiple angles while always maintaining skepticism and inquiry.

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle topics to give a 360-degree view on this vast, ever-changing industry.

Dynamic Creative: The Perfect Gift for Every Campaign on Your List

Posted in Audience Series, Most Recent

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic to give a 360-degree view on this vast, ever-changing industry.

Young Lee
By Young Lee

Being in the midst of the holidays, consumers are hustling to find the perfect personalized gift during this shopping season. The same is also true for marketers and media buyers, where they strive to deliver the most relevant and personalized ad to their target audiences.

Using PlaceIQ’s technology, marketers are able to build sophisticated audiences based upon real-world behavior, and then leverage the context of time and location to deliver the most relevant ad; however, this is only half of the puzzle. The other half is the mobile creative. Many marketers execute campaigns with a “one creative fits all” approach across the targeting tactics of their mobile campaigns. However, it’s essential that one must understand that mobile is different from desktop because of the location-awareness of mobile devices. Delivering audience-aware mobile creative can have interesting results.

Audience-Aware Creative

When building out the mobile campaign creative, it is important to consider the audience, location, and time of day to give that personalized touch and resonate with the end user. Choosing the right creative is in itself a daunting and complicated task. Below are some suggestions for improving mobile creative, gleaned from what I’ve observed with working with marketers.

  1. Lower- Versus Upper-Funnel Tactics

    Targeting users in the lower funnel, like in-store or at the point of purchase, should have messaging related to features, price-points, and product descriptions. Users in the upper funnel should have more general brand messaging.

  2. Demographics

    Marketers should consider the demographics of their users. For example: Is the targeting tactic reaching price-conscious or affluent users? Affluent users are likely price-insensitive, so messaging should include features and benefits.

  3. Location

    Because the duo of location and time is a large indicator of intent, taking the combination into account for creative messaging is also important. Is your targeted audience interested in music? Have they been to a music venue recently? Were they there at night, perhaps when a concert could have taken place? This user will likely respond to messaging that includes a music theme.

Measurement

For any creative execution, it’s always important to have some sort of measurement to track the results. While click-through rate and interaction rates are the standard metrics of performance in display and mobile campaigns, PlaceIQ’s Place Visit Rate™ (PVR™) is a strong indicator of performance when it comes to branding campaigns because it can help marketers determine the true impact of their ad on in-store visits.

An example of an early success of utilizing “audience-aware” creative was a recent mobile campaign with an auto advertiser driving consumers to their dealerships. This campaign tested placements using generic creative against dynamic creative, in which the user received mobile ads that changed messaging and background depending on a user’s location. Initial results are showing positive signs, where users exposed to dynamic ads are about 2 times more likely to visit target destinations when compared to the other targeting tactics.

Choosing the right creative (or the right gift) is always a difficult task. There is never a silver bullet, and it will require multiple rounds of trial and error. However, learning about your target audiences and tailoring the creative based on this data can go a long way to connecting your brand to your audience.

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic to give a 360-degree view on this vast, ever-changing industry.

What We’ve Learned: Navigating an Event-Driven Mobile Campaign

Posted in Audience Series, Blog

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic to give a 360-degree view on this vast, ever-changing industry.

Joeseph Ranzenbach
By Joseph Ranzenbach

PlaceIQ CEO Duncan McCall has long said that “location is the biggest indicator of intent since search,” and both brands and advertisers have seemed to agree in kind. Mobile, the media format that enables the greatest degree of accuracy in identifying a user’s location, has witnessed a growth in advertising dollars that outpaces every other media format, and much of this spend is going to location-based targeting.

And why not? Not only are there more Android devices activated each day than there are babies born, but users are spending increasing amounts of time on their newfound electronic lifeblood, which is rarely more than three feet away from them. Add in granular location data on each impression, combine that with a sophisticated understanding of the space and time in which consumers travel, and you have an incredible platform for audience development, targeting, insight and attribution.

But in order to build and execute on a successful campaign, particularly for the impending phenomenon of holidays over the coming weeks, it is imperative to develop a well-planned, event-based advertising strategy.

The Challenges of Event-Driven Targeting in Mobile

Targeting a successful location-based advertising campaign on a day like Black Friday, Cyber Monday, or any of the hectic shopping days before Christmas can prove to create an enthralling set of problems to solve. In our experiences running event-driven mobile campaigns in the past, including a particularly relevant, large-scale campaign that targeted in-store customers on Black Friday 2012, we’ve learned to fine-tune our execution strategies to account for a few potential challenges:

  1. Supply Scarcity

    Whether you’re a big box retailer with 1,000 brick-and-mortar locations across the country or a regional consortium of auto dealers, it’s important to note that there is a limit on the market of available impressions out there for your in-store customers. (Khoa Pham wrote an excellent post on Fermi problems and ad targeting).

  2. Increased Competition

    Increases in consumer purchase intent on days like Black Friday are met with increases in advertising demand, meaning that the competition for consumer attention on limited available impressions is much higher than usual and, despite increased shopping activities, supply will not rise to meet demand in many cases.

  3. Scalable Infrastructure

    The increasing rush of ad demand to RTB, which has consistently outpaced research firm projections in recent years, has led to an increased supply in ad impressions and, consequently, an increased requirement for ad buyers to develop scalable infrastructure. The increase in consumer activity and competition for impressions brought on by holidays and large-scale events makes scalable infrastructure even more important. In order to capitalize on as many of those limited, relevant available impressions and beat out competition for them, it’s imperative to support seeing as much of the pool as possible. Unfortunately, due to the innate scarcity of addressing in-store audiences, the law of diminishing returns applies to supporting increasing pools of inventory and infrastructure.

  4. Atypical Behavioral Patterns

    Large-scale events effect consumer behavior and require algorithmic adaptation and iteration. What works on a typical Friday will not likely produce the same levels of accuracy or success on Black Friday or the Super Bowl when it comes to targeting, analytics, and performance (Rachit Srivastava wrote a great post on this topic).

Making Location Scale & Constructing Successful Campaigns

So given the known limitations, how do you make location scale to most effectively reach your audience and meet your campaign goals?

  1. Build a Strategic Audience Portfolio

    In building your campaign, it’s important to develop a targeting portfolio that not only helps you to achieve your goals, but also hedges your risk and offers opportunities for some home runs. For instance, while targeting consumers that are in your brick-and-mortar locations is an exceptional component to any campaign, it shouldn’t be your entire campaign, as it may limit your scale and reach. As supplemental audiences in your portfolio, why not try conquesting those who have visited a competitor’s location in the past and also targeting locations that consumers are likely to visit before reaching your stores?

    Additionally, it’s worth noting that growth in mobile commerce is outpacing both e-commerce and in-store sales. Why not build that into your campaign strategy as well? Given that not all consumers will be doing their shopping in-store this year, adding line items for in and out-of-home audiences with affinities for your products can help capture mobile or online commerce in addition to helping drive brick-and-mortar traffic.

  2. Utilize Data Intelligence, Not Data Sets

    There are a lot of flawed and incorrectly attributed location data sets out there and unfortunately a lot of folks are using the same ones. When dealing with large amounts of spend, especially over short periods of time, invest in a partner who invests their own time and resources in data aggregation, intelligence, quality, and analytics. There’s a reason why Apple’s initial mapping foray met so much criticism – location is a lot harder than it seems.

  3. Plan Ahead & Close the Feedback Loop

    At PlaceIQ, we’ve plotted billions of points of information against our patented location analysis platform to derive an intuitive, audience based understanding of the world around us. In preparation for large-scale, event-based campaigns, we reference previous campaigns and data sets with billions of data points and comparable conditions to adapt our expectations, rather than, say, a panel of a few thousand users (Extra Credit Reading: Why We Need to Do Better than Panels, Focus Groups and Surveys).

The holidays can be a very stressful time for many brands and advertisers, but they really don’t have to be. Significant scale and profitable results can be achieved by intelligently targeting your ad spend through strategically built audience compilations and being flexible with iteration. When it comes to event-driven targeting in mobile, thoughtful planning and leaning on data-driven insights can make the difference between celebrating an innovative and successful campaign and standing still while your competitor does.

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic to give a 360-degree view on this vast, ever-changing industry.

Black Friday: The Ultimate Test

Posted in Audience Series, Blog

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic to give a 360-degree view on this vast, ever-changing industry. The following article was featured as an AdExchanger “Data-Driven Thinking” column.

Rachit Srivastava
By Rachit Srivastava

Thanksgiving is around the corner, meaning great food, family and, of course, Black Friday.

Mobile ad targeting and strategy have advanced significantly this year, and mobile marketers are reaping the rewards of proven, solid algorithms and enjoying consistent uplift in success metrics. They should be feeling pretty comfortable with being able to target audiences efficiently on this crazy shopping weekend, right? Why would their algorithms fail them now?

Black Friday is like no other time of the year, and it may require marketers to step out of their comfort zones. This is no time for autopilot. If you want to pass the ultimate test, preparation is key.

Major Challenge 1: Infrastructure

On Black Friday, consumers are more inclined than ever to shop, and advertisers want to take extra advantage of this. This means that overall there will be more competition for targeted ad impressions served that weekend than on other days of the year. Advertisers will funnel an increasing number of ads towards these limited impressions in order to cash in on the intent of consumers and channel them toward stores.

To read the rest of this article, please visit Adexchanger.com.

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic to give a 360-degree view on this vast, ever-changing industry.

6 Reasons Why Data Science is an Art

Posted in Audience Series, Blog

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic each week to give a 360-degree view on this vast, ever-changing industry.

Juan Huerta
By Juan Huerta

If Michelangelo were alive today, what would he do for a living?

He would be a data scientist, of course.

Sure, crunching massive data sets on multi-thousand core clusters using algorithms that were once the exclusive domain of the scientific elites might not seem like an obvious career choice for the famous Caprese maestro, but I firmly believe that he would make quite the data scientist.

Here’s why: Behind the name, data science is a transformative craft, which shares a number of similarities with art.

www.freakingnews.com

www.freakingnews.com

  1. Transformative Synergism: In both disciplines, the final total is much greater than the sum of the inputs. In data science our prime raw material (data) goes through the process of transformation through cleansing, parsing, and normalization, followed by iterative optimizations where incremental value is distilled out of the raw material, and where a final result emerges in the form of new information. As point in case, imagine the workshop of Antonio Stradivari systematically transforming a generic block of wood into a unique piece of acoustic excellence.
  2. Technique and Apprenticeship: Art, like data, is based on mastery of technique. A painter takes pride on technical skill the same way a data scientist takes pride on the techniques he or she can bring to the table. And, in both cases, while fundamentals are learned through formal education, true technical mastery is only developed over periods of time and through hands-on experience. Data organizations are essentially workshops where technique is constantly emphasized, transferred, and codified in the form of best practices, intellectual and proprietary information, and techniques.
  3. Experimentalism and Innovation: As piano technology evolved during the 18th and 19th century, composers of the era quickly assimilated the innovations to the instrument and dutifully reflected these into their music. One can imagine the excitement of a young Beethoven when noticing the expressive range of the new pianos of the time. Data scientists have the same curiosity and adventurous spirit as we constantly assess every new technology that promises to produce improved outcomes and workflows.
  4. Creativity, Imagination and Hacker-spirit: Anybody who has worked on code, models, algorithms, or building data pipelines knows that in our line of work, inspiration and creativity are crucial. Sure, we still need the proverbial 99% perspiration, but if you haven’t got the inspiration, no amount of perspiration is going to get you out of your predicament. An online definition of “hacker” refers to a person “who enjoys the intellectual challenge of creatively overcoming or circumventing limitation.” Creativity is a strategic weapon in the hacker’s arsenal. One story tells us of Paganini occasionally breaking his violin strings during performances to demonstrate his virtuosity. Creatively overcoming these self-imposed limitations earns Paganini the honorary title of “hacker.”
  5. Specialization and Differentiation: Nobody disputes that Leonardo’s art is distinctive. Mozart never said “my music is awesome because it sounds just like Bach’s.” Nobody mixes their Dali’s with their Picasso’s. Every data organization strives to achieve a differentiation in technique and approach in order to produce characteristically distinctive results. Like artists, top data science organizations typically operate within their own niches excellence, reflecting the way we think of and approach a problem, our domain of expertise, as well as the resources we leverage and techniques we bring to the table.
  6. Persistence and Detail-orientation: Those of us who have seen architectural marvels like La Alhambra in Spain have inevitably wondered how many thousands, if not millions, of man-hours were spent creating such extensive and magnificent works. Likewise, data science is done through careful persistence and fastidious attention to detail. The true data scientist will understand that the difference between a tight model and a sloppy one is the belief that no detail is too small to matter and that in the end, this obsessive persistence is what makes all the difference.

We can go on listing many other similarities, but the ones mentioned should illustrate the main parallels between data science and art creation. I have no doubt a latter-day Michelangelo would definitely relinquish his chisel and mallet for some Hadoop and NoSQL.

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic each week to give a 360-degree view on this vast, ever-changing industry.

The Year of Mobile?

Posted in Audience Series, Blog

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic each week to give a 360-degree view on this vast, ever-changing industry.

Duncan McCall
By Duncan McCall

We’re all amusingly familiar with the seemingly annual ritual that takes place with various analysts proclaiming that the upcoming year will be the “Year of Mobile.”

This has always struck me as analogous to saying it’s the “year of the car” or the “year of the computer.”

It bugged me that no one really took this statement to task, so I gave a presentation recently titled “The Year of Mobile,” and when I started really thinking about this concept, it became quite an interesting exercise.

So what year could actually be considered the mythical “Year of Mobile” if you were to actually try and identify one?

Maybe it was 1983, when Motorola launched the first mobile phone: the DynaTAC (forever immortalized by Gordan Gekko). That had to be the year of mobile…right?

Slide1

Or perhaps 1989, when the Motorola StarTAC came out, and every sales rep worth their salt was flashing off their slick new clamshell mobile.

No? Well then what about 1993, when Nokia launched their first real GSM phone, complete with a calendar and the oh-so additive game of Snake.

If not then it had to be 1999. Blackberry took email on a phone mainstream with huge success – a watershed moment if there ever was one.

What about 2007? This had to be it. iPhone and Android ushered in the smartphone with what was to be the fastest consumer adoption of any consumer technology in history.

Slide6

Then of course it dawned on me: There was no “Year of Mobile.” We are simply living in the Age Of Mobile.

As we sit now with this so-called smart phone revolution well underway, we have these always-on, always-near, location-aware, connected devices that we check 40 times a day.

We’ve all seen first-hand how these devices have changed our daily lives.

They’ve changed the way we consume news and take and share photos. They’ve rendered paper maps almost obsolete. Ultimately, they have changed the way we actually behave, communicate and interact with one another.

They are also starting to change the way that brands and advertisers communicate with their customers and learn how their products are interacted with in the real world. They may even provide a platform to realizing the long, unfulfilled dream of one-to-one marketing.

Even more, we’re still in the very early days of this revolution. We’ve only just passed majority usage in the US, and with wearables, indoor location, health and fitness, mobile payments – all barely getting started, but along with many other mobile integrated technologies – poised to have a tremendously disruptive and transformative effect on our lives.

Analysts search for some magical metric in a short time horizon to identify that mobile has finally arrived. I think that ultimately we are in the middle of a technology phase that will be measured in decades and will continue to morph and change and defy simple labels or even accurate measurement.

So here’s to embracing the fact that we’re not in the “Year of Mobile.” We’re simply living in the Age of Mobile.
Slide22Y

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic each week to give a 360-degree view on this vast, ever-changing industry.

Why We Need to Do Better than Panels, Focus Groups and Surveys

Posted in Audience Series, Blog

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic each week to give a 360-degree view on this vast, ever-changing industry.

Niladri Batabyal
By Niladri Batabyal

In perhaps the most inspiring development in the literary world in recent memory, a survey of Millennials showed that 100% of respondents still read books on a daily basis. Though works of nonfiction outpaced “reading for pleasure” by almost 2 to 1, a whopping 33% of the population is engrossed in A Game of Thrones.

Though the remake on HBO has received critical acclaim and a very healthy fan following, the results of this carefully selected focus group prove that Americans still appreciate those special moments of solitude in their lives when the simplest words on a page can transcend their very being into a new reality.

Skeptical?

Well, you should be. The “study” above was conducted with a group of my colleagues over coffee. Though there is nothing scientific about that particular selection process, a market starved for consumer insights must often be reminded of the cautionary tales in Darrell Huff’s How to Lie with Statistics.

The lessons there within seem as true today as they were back in 1954 when this classic was first published.

Through flowing numbers, beautiful graphs, and flowery words dressed as conclusions, many a consumer study may appear sound. Although the results should be scrutinized from top to bottom, often the biggest holes can be found by looking right at the inception and asking one simple question: Who was in the sample?

www.treehugger.com

www.treehugger.com

A History of Surveys

When Arthur Nielsen started his radio rating service in 1942, you can imagine the Great American Family sitting around their radio after dinner, listening to a small group of national broadcast networks all competing for the national set of eyes and ears.

At that point, it’s not that far-fetched to believe that the viewing habits and interests of the Great American Family could be explained through a magnifying glass on a small sample of households.

Over time, that Great American Family has changed, along with their behaviors, intents and very nature. Mom and Dad are driven by different factors. The kids live in a world of their own. Comparing the Midwest to the Northeast may as well be like looking across oceans.

Time for Change

When you consider the incredible technological advances that have been made in media over the last half century – from television to the internet and now on to mobile – the sheer scale of consumption and variety of choices has grown exponentially for the American consumer. In attempting to understand the who, what, when, where and how of their journeys, do we as brands and marketers really have an exponentially better understanding of consumers today than we had 50 years ago?

Moreover, in a constantly changing, ever more connected world where global trends are overshadowed by local influences, should we still extrapolate the results of, say, a focus group of 100 to the general population of 300 million?

When I sit back and think about my own decisions on a daily basis, the paths that I take are immensely varied. I can watch one of 500 channels on television. I can choose to experience just about any type of cuisine from around the world. I can get my news from a near infinite number of local and social sources around me. On a Friday night, I can go to the opera, the symphony, a rock show or chill out to some smooth jazz.

So, as a marketer, if you want to connect with me, understand me, and reach me with the right message at the right moment in my life, are you seriously considering using the experiences and habits of a college student in New Mexico who fills out a survey for a little extra beer money to mirror my own in that all important 18-34, male demographic?

Now, chances are we both like beer, but that’s probably where most of the comparisons end.

How to do Better

Let’s stop taking the easy way out. Rather than just repurposing dated methodologies invented half a century ago, it’s time that we change the entire conversation around consumer intelligence.

Easier said than done, you may say. A paradigm shift is rarely straightforward. When the goal is to understand the new reality, the road to nirvana becomes even more daunting.

Being responsible for innovation here at PlaceIQ, I take this task very seriously and believe we are on to some major disruptions. If we start with the Who and attempt to extend our breadth and depth of understanding based on reality, we can let the past, present and future actions of the Who paint a complete picture.

Consider this:

  1. Let’s start with the premise that location is the greatest indicator of consumer intent since search. Those little devices that we carry around with us everywhere provide the perfect backbone for our new reality. At scale, we begin to see patterns emerge from hundreds of millions of devices emitting hundreds of billions of signals every day.
  2. We can connect data across multiple channels and overlay it on our foundation based on location. Desktop cookies, television and radio viewership habits, past purchase behavior, social trends – the list goes on.
  3. Finally, we bridge the divide between the online and offline worlds. This is where where the old surveys or focus groups can still provide some value by plugging into the new. The CRM systems and offline consumption behaviors that brands have leveraged for decades can close our proverbial loop and complete our view of the consumer journey.

I hope this post has addressed why, as an industry, we need to do better. The framework above should provide a glimpse of how we are already on the right track.

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic each week to give a 360-degree view on this vast, ever-changing industry.

The Optimal Audience Portfolio

Posted in Audience Series, Blog

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic each week to give a 360-degree view on this vast, ever-changing industry.

Rachit Srivastava
By Rachit Srivastava

Do you manage a stock portfolio? Have you ever tried to create an optimal portfolio?

Interestingly, creating an optimal portfolio and deciding on an audience selection for a campaign are more similar than they may seem at first glance.

Reducing Unsystematic Risk

Creating an optimal portfolio of stocks can be a tricky task. One must diversify risk by choosing stocks that are uncorrelated or, ideally, anti-correlated. In other words, if a stock in the portfolio goes down in value, then there should be another stock that characteristically would counter this downward movement by going up. This attempts to cover what financial analysts call “unsystematic risk.”

Audience selection for an ad campaign can be thought of in a similar way. Consider the following task: Serve a given number of impressions and maximize the number of impressions served to a person belonging to at least one audience in a basket of audiences.

When constructing this basket of audiences, one must be very careful not to include audiences that tend to co-occur together. What do we mean by co-occurrence?

A co-occurrence at the location level means that certain audiences will either all exist together or not exist at all at a given location.

Likewise, at the unique identifier level, this term means that if a person belongs to one audience, then they are very likely to belong to another audience (and vice versa).

We want to avoid this because when serving an impression, if you’re not reaching one audience the hope is to reach another. We don’t want to choose audiences such that missing one means we are also missing the other.

Consider a luxury car owner and a resident of an affluent neighborhood. Now say we are targeting these audiences based on location. Luxury car owners will tend to live in affluent neighborhoods and residents of affluent neighborhoods will tend to own luxury cars. If we create an audience basket with both of these sub-groups, there is a low likelihood of hitting either group if the other is missed.

This unsystematic risk can be taken care of by choosing segments of populations that tend not to co-occur. In the above example, a better choice may be to couple luxury car owners and fashion retailers. Fashion retails will generally not occur in the same location where luxury car owners are living. This will allow us to increase the reach of our targeting.

The Liquidity Factor

In another scenario, consider wanting to sell out of a company due to a bearish outlook about the success of a new product. The market, on the other hand, is not trading much of this security because people are very uncertain about whether the product will catch on.

If I believe that the fair price of the stock is $100, but nobody is willing to buy it for over $97, I will either have to decrease my offer or risk it not being filled. If I decrease the price, I will have paid a transaction cost of the difference, and this transaction represents the liquidity of the stock.

Interestingly, we have an analogous situation when trying to serve the target number of impressions to our basket of audiences. Consider a scenario in which we can confidently assert that if an impression is served at a certain location, then the person likely belongs to a certain audience: 18-25-year-old Australians, for example. Even though there might be locations such as an Australian sports bar that have a high ratio of such audiences, the number of people and total number of such bars will not be very high.

This means that having a basket of such audiences will make it hard for our impressions to be served. Sure, we could target random audiences with a lower anticipated CTR, but this would be parallel to lowering the price of the stock, just to ensure the order is filled.

The illiquidity of “low reach audiences” represents a major challenge in mobile ad targeting due to the tradeoff between high fidelity and reach. Generally, a high-fidelity audience will have a low reach and we must compromise on some level of fidelity to gain reach. In the world of stocks, we have a tradeoff between taking advantage of imperfect information and the liquidity of stocks.

Implications for Wall Street?

Does this mean that people who are good with stock trading will be good at mobile ad targeting? Not necessarily.

One obvious barrier is that domain knowledge – like with most things – is very important. A more subtle factor is that the outcome of stock trading strategy can be objectively measured with a metric that is consistent with the explicit goal (i.e. profit and loss). Things are not as objective in the mobile advertising world where performance is measured by the highly contentious CTR.

Despite these challenges, the mobile advertising market is evolving with metrics like PVR proving to be very valuable. As more such metrics are created, these financial optimization tools will be even more applicable.

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic each week to give a 360-degree view on this vast, ever-changing industry.

Data Science Sanity Check: How to Ask the Right Question

Posted in Audience Series, Blog

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic each week to give a 360-degree view on this vast, ever-changing industry. The following article was featured as an AdExchanger “Data-Driven Thinking” column.

Khoa Pham
By Khoa Pham

How many golf balls can you fit in a 747?

This simple, albeit somewhat ridiculous question digs right into the heart of data analysis and is the bane of recent college grads looking for employment in any quantitative field.

How many hairs are there on a human head? How many piano tuners are there in the city of Chicago? How many words have been spoken since the beginning of time?

These questions test the way we go from a series of assumptions about things we have limited information about to coming up with an answer we can’t immediately test.

This type of estimation, called a Fermi question, is often used in engineering and sciences to scope a problem before attempting to build a complex model to derive a more precise answer. This more precise answer is often not needed and expending time and energy to calculate the answer can hurt the bottom line of the business.

This couldn’t be truer in data science, where money can disappear into black holes of “data analysis.” Having complex systems in place to compute forecasts and project inventory does not mean they should be used for every question that arises.

The influx of interest in mobile advertising, coupled with the recent infusion of “Big Data” into the ecosystem, has brought these types of Fermi questions to the forefront of an analyst’s day-to-day job.

To read the rest of this article, please visit Adexchanger.com.

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic each week to give a 360-degree view on this vast, ever-changing industry.

The Journey is More Valuable than the Destination

Posted in Audience Series, Blog

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic each week to give a 360-degree view on this vast, ever-changing industry.

Brian Stoller
By Brian Stoller

“The journey is more valuable than the destination”

This Chinese proverb, dating back to 230 BC, has never had more relevance than in today’s modern mobile ecosystem. Too many marketers are considering location as a shopper marketing opportunity, trying to open a dialog with the consumer while they are in-store.

Although there is tremendous value in targeting these destinations, the journey these mobile devices make over time is far more valuable data than the contextual location of a user at a given moment.

We’ve all heard the same mantra for the last 10 years: always within reach, most personal of devices, et cetera. So here is a device that travels with every individual, and is unique to that person.  Unlock the journeys that these devices make, and a world of new audience targeting becomes available to the modern mobile marketer.

Consider this: A B2B marketer seeking to target a business decision-maker could target airports, attempting to reach a business traveler. This is not a bad strategy; however, the amount of waste in targeting tourists, retirees, and other general travelers is equivalent to buying a home page takeover on any large portal or newspaper.

The better solution is to consider not only the airport, but the frequency of travel through the airport as the identifier of a heavy business traveler. By identifying an audience as people present in airports, say, three or more times in a month, traveling between Monday and Friday, we’ve eliminated the casual traveler.

Now take those same devices and apply additional data filters: owns a luxury automobile, lives in an affluent neighborhood, works in a white collar zone, and so on. The scale of these highly filtered location-defined audiences drops significantly, but the end result is a highly defined senior business decision maker.

I hear mobile marketers and advertisers continuously stating that we are in an “early stage” of development; that mobile is still a nascent marketing entity. The reality is that most marketers have approached mobile much in the same way they approached other digital and broadcast mediums before, using contextual targeting to identify an audience and seeking a solution for the lack of cookies.

By attempting to apply mechanics from other channels to mobile, the process of trial and error has hindered our ability to rapidly embrace mobile’s unique marketing characteristics. Instead, we need to re-evaluate everything and look to those differentiators between mobile and other digital mediums.

The largest element of disruption for mobile comes from the fact that this device is, in fact, “mobile.” We’ve failed to identify that the medium’s name itself defines the very asset we should seek to embrace. The true mobile marketing sage will consider how the journeys these devices take can create more valuable lifestyles segments and audiences, and avoid an emphasis on the destination itself.

PlaceIQ’s Audience Series sets out to highlight the importance of segments in the advertising world. As the pioneer of mobile’s application to location intelligence, and leaders in the mobile audience field, PlaceIQ has the knowledge you need. Key audience experts from each PIQ department — from engineering, to data science, to sales — will tackle a new topic each week to give a 360-degree view on this vast, ever-changing industry.