CASE STUDY: Gaining Sales Lift Through Attribution Modeling
Closing the Tracking Loop for GLOCK Drives Optimized Media Mix
Over 65% of US law enforcement agencies rely on GLOCK pistols. Also highly respected by civilians, they reach a broad range of customers for personal defense, sport shooting and hunting needs. Media Two handles all media buying responsibilities including digital, traditional and sponsorships.
In the firearms industry, the ability to track point-of-purchase is extremely difficult. The supply chain relies on a small number of distributors that sell to dealers. Pistol sales at the dealer level are also rarely tracked beyond federal background checks. The manufacturers’ ability to collect customer data is limited – firewalled by both the dealers and the consumers.
GLOCK approached Media Two with the need to establish point-of-sale tracking across all of their various marketing strategies. Their desire was to gain a holistic understanding of what role each of their marketing channels – paid, owned and earned –played in driving sales. This level of insight would be used to more effectively deploy resources and budgets – especially with the launch of new product lines.
GLOCK’s marketing footprint consisted of paid media across digital, television, print and sponsorships platforms. Adding to this was a strong social presence through owned channels, very heavy exposure through public relations efforts in industry publications, as well as high traffic volume to their website.
Within the supply chain, only three sets of data were available – sales to distributors, distributor sell-through volume to dealers and registered product warranties. Actual consumer sales data from dealers would not be available.
Our specific task was to connect the data points in such a way that we could effectively analyze and optimize the path to purchase. Media Two would handle all media buying responsibilities as well as data analysis, while GLOCK’s internal marketing team would be responsible for the public relations, social media and site optimizations.
Establishing an extrapolation model
At the core of our approach was the establishment of a full-funnel attribution model. Warranty activations, as a post-purchase action, effectively gave us a conversion metric upon which to base our findings. With the distributor sales and sell-through data available, using the warranty data allowed us to apply an extrapolation model to represent total sales from the manufacturer.
Additionally, this would serve as a point to collect valuable insights on the specific types of purchases being made. Tracking data points such as gender, first-time buyers, needs driving purchase, and B2B versus B2C sales allowed for segmentation of the attribution model by each data point. For example, segmenting law enforcement purchases gave us a very different model by which to optimize our marketing against compared to how we may target and optimize for first-time buyers.
In order to effectively analyze the distribution path, it was also necessary to gain insight into the path length from shipping to final in-store transactions. Working closely with GLOCK’s marketing team, we helped to create an in-box promotion incentivizing the completion of a warranty. By doing this, we could track the specific shipment dates of product from the manufacturing facility to the registration of the warranty and actual sales date. Understanding these timeframes would play a key role in determining frequency and remarketing strategies within the media buys
Finally, working with various media partners, we implemented attribution tracking across our media buys as well as establishing the necessary data feeds from relevant sources. Television log files, site analytics, print tracking, social feeds and ad serving were all tied to the platform in order to centralize the data.
Putting cross-channel data to work
With tracking in place, we began to monitor the attribution data to determine when statistical significance was achieved. In order to do this, the specific KPI we tracked was lag time from first exposure to conversion. Once the lag time stabilized, that was an indication the model was effectively including all data necessary to analyze path-to-purchase.
By analyzing the various marketing exposures we were able to determine our effective window for driving conversions was a 25 day average with an optimal frequency of 15 advertising impressions. Additional insights drove optimizations shifting print into television, and shifts in digital strategy focused on non-endemic prospecting to women and first-time buyers as well as latency and frequency capping for remarketing.
Of all sources tracked for the year, paid media was credited with driving 45% of total product sales and we saw impressive improvements in the primary campaign KPIs:
- 42% lift in brand engagements
- 93% lift in attributed sales
- 59% improvement in cost-per-unit-sold
As we increased our prospecting strategies in digital and TV, we also noted a lift in the role SEO and organic sources played in the attribution model. This was an indication that paid media was also driving engagement with other marketing channels.
Qualitatively, our media strategies aimed at reaching audiences outside of the endemic publications was also very well received by consumers. Shortly after launching these efforts the brand received several comments from social media fans supporting our efforts:
The overall approach and results have far exceeded GLOCK’s expectations, and the model continues to evolve with their
ongoing marketing efforts.