This post kicks off our annual Predictions series, underwritten this year by Freckle IoT, a first-party data company that solves in-store attribution for brands and agencies. This post offers a preview of Freckle IoT’s plans for 2017. Catch up with all the 2017 forecasts here.
By the end of 2016, annual digital ad spending will exceed $72 billion—representing over 36% of total ad spending in the market. While this number has grown exponentially, the ability to determine how effective advertising has been in driving customers to a desired location has not. This is a problem, considering that 90% of all purchases still take place in a physical store. Our industry has made great strides in solving viewability and fraud issues, but in-store attribution remains unresolved. However, I believe that by this time next year the majority of mobile campaigns will contain an attribution metric. Why? Demand for data and transparency.
As brands now demand all ads be viewable and free of fraud, the natural iteration is to then demand to know that they work. The companies who prove that their media is better at driving in-store visits will reap the benefits from an industry that has longed for this metric. Measurement will not come from the vendor itself but rather from third-party measurement firms decoupled both from the buying and selling of advertising and from the platforms on which the media runs. The days of self-measurement are over.
Having previously founded a company that built a mobile DSP as well as the world’s first mobile futures market, I am well versed in the limitations of bid stream stream data; the current methods of attribution; and the buying and selling of mobile advertising. Recognizing these limitations led me to the creation of Freckle IoT, a first-party mobile data company focused on the measurement of media, not the buying and selling of it.
Today there are multiple references to attribution being tossed around in the market but not all attribution is the same. For brevity I will focus on in-store attribution. In-store attribution is the process of understanding which consumers made a visit to a physical location as a result of seeing an advertisement. Many in the proximity space have made claims that they are able to measure attribution via bid-stream, triangulation and other tactics, but all are flawed. In reality, without some physical tie-in to the store such as a sensor, a beacon, a coupon or a credit card receipt, this is all probabilistic, which is another word for a guess. In media, have we not learned that deterministic ID’s trump those that are probabilistic? Suggesting you can measure in-store attribution without a one-to-one match is…generous.
As brands look towards attribution in 2017, they must consider a few other key points. First, there is a lot more data floating around this ecosystem than most realize. This is not a cottage industry and is reminiscent of programmatic in the early days—fast and loose. On top of this, there are real issues around the origin and the fidelity of this data.
Here is one guiding principle: If you don’t know the origin of the data, discard it. By default, this removes third-party bid stream location data from the attribution equation. Not only is bid-stream location data notoriously inaccurate, its limitations are compounded by its inability to measure core in-store metrics, such as enter and exit events or dwell time, due to its dependency of a user having to be in a application session. Solutions which allow for in-store data collection irrespective of session use, such as an SDK, solve this.
Regarding fidelity, the guiding principle is that not all data points are equal. The range of fidelity is anywhere from 1 to 40-plus data points per day, depending on the solution and the source of the data. The more data points, the higher the fidelity and the more accurate. If one looks to measure in-store metrics, knowing the entire path to purchase requires more data—not something that can be accomplished via a singular data point.
As in-store attribution matures this year, third-party measurement of high-fidelity first-party data will quickly rise to become the default. As this takes place, all brands should be looking in the mirror and asking this question: “How effective is my media in driving people into a location and what is my plan to measure it?”
You better have a good verifiable answer.
Neil Sweeney is the founder and CEO of Freckle IoT, a first party data company that solves in store attribution for brands and agencies.