The Cocktail Discovery Model brilliantly elevated old-fashioned keyword research and taking it from a chore to a business intelligence and business change enabler. By understanding the true “Voice of the Consumer” we are able to create effective engagement opportunities that contribute directly to bottom-line business growth.

Enhancing time-consuming human segmenting with Artificial Intelligence and Data Visualization enabled faster classification and refinement of the data into useable clusters. Integrating ongoing performance and creative engagement data further enhanced the process by isolating content that better matched the actual interest of the consumer, fostering deeper engagement and interest. This created a repeatable process that generated incremental insights as the dataset grew.

The initial data analysis identified over 500 new content opportunities and changes to current assets. Most importantly uncovered the “inspiration triggers” consumers use to discover new drinks for specific occasions, by colour, origination, or even glass types.

These insights have created a significant number of new opportunities for the brand to generate consumer-specific content for not only the website but in social media, how-to videos and the functionality of the cocktail app.

Due to scale, we focused initial efforts on the “Drink Curious” cluster of 53 cocktail attributes and 107 occasions. The plan was to continue to refine a deeper understanding of these consumer interests and make appropriate content adjustments.

A key example of the insights was a test to adapt cocktail images for mobile devices from the traditional vertical to horizontal. This relatively simple creative change increased clicks on mobile recipe cards by 67% which was a major contributor to our goal attainment.

As our confidence in the data and our ability to refine the insights increased, we expanded out further into the 250 different clusters looking for additional targeting and product insights. In total, we integrated data from 43 different sources expanding the dataset to over 10 million records across 1,000 variables. This has included launching nutritional information as well as expanding the product FAQ and sharing these insights with e-commerce channel partners.

This data-driven innovation fundamentally changed the way the brand thinks about consumers, content that really generates engagement, and creates a sustainable competitive advantage that increases mindshare and ultimately revenue.

This program exceeded established objectives beyond our expectations. Initially visits increased 24% with mobile traffic specifically increasing by 77% with a total increase of 125%.

Consumer experience improved, demonstrated by page views increasing from 1.3 to 3.6 per visit. Page views for “Cocktail Curious” clusters achieved 8.3 with a significant increase in eCRM signups far exceeding the goal of 50%.

An excellent example of business impact is showcased in deeper analysis of the popular phrase “drinks with triple sec.” The optimization analysis identified consumers wanted a list of drinks with this ingredient. The existing content described triple sec and said it was used in popular cocktails. Revising the description to include recognizable cocktail names, interaction went from 1.2% CTR to 16.9% driving incremental visits and a 10x increase in signups. This enhancement was deployed across the entire “drinks with” content portfolio.

The data mining and refinement outlined above enabled consumers to discover new and interesting cocktails but, more importantly, helped reinforce Absolut as the vodka brand to consider when looking for interesting cocktails. The increased signups enable Absolut to retarget this audience multiple times in the context of their initial interests, expanded interests, fostering engagement and brand loyalty.

The data-driven Cocktail Discovery Model is an extensive analysis and creative optimization protocol that leverages searcher interest data to understand how consumers find inspiration for new cocktail ideas.

Using human and artificial intelligence, converted 10 million data points into 1.5 million cocktail specific queries, further segmented into 250 category clusters. These clusters revealed over 500 new content opportunities, when implemented and refined, resulted in significant increases in visits, eCRM signups, but more importantly, a level of engagement that is not possible, or affordable, with other forms of advertising.

Data gathering started with the aggregation of multiple internal datasets starting with the drink recipe database. This dataset had over 8,000 cocktail recipes. Of specific value were the name and ingredients of each cocktail. The next step was integrating social media and search queries that from various websites followed by 3rd party lists from the spirits and beverage industry.

When text mining and visualization were applied three core clusters initially emerged. The first was related to specific cocktails and how to make them. The second consisted of various holidays and other occasions such as Halloween cocktails. The third cluster was the goldmine with elements like “drinks with x,” “frozen drinks” and “red cocktails” which resulted in the further manual and automated expansion of these terms.

Once we had a significant dataset we integrated Google Search Volume which told us how many searches each variation had in a month. This data point helped us aggregate concepts, order clusters and content types then prioritize our targets based on searcher interest levels. Starting with the clusters and phrases with the most searches, and the highest probability for engagement, we created relevant content and pushed it to the search engines.

The brand had been given a goal of significantly increasing eCRM signups. Analysis of current signups indicated organic search, but more specifically “Drink Curious” searchers, those looking for drink ideas based on spirit, occasions or ingredients had significantly higher eCRM signup rate than generic searchers.

The brief was simple, use data to find these elusive “Drink Curious” searchers.

What words do they use?

What attracts them and encourages interaction?

What will prompt them to sign up for the Drink of the week email?

The objectives were clear.

Objective 1 – Find the Drink Curious in data then,

Objective 2 – Identify and create content that attracts, engages and retains them, then

Objective 3 – Create a repeatable process that keeps delivering Objective 2, then

Objective 4 – Leverage this process to help capture at least 50% of eCRM goal.