Brand Battles in Depth: Looking at Starbucks vs. Dunkin’ Donuts

Share this:

Last week Street Fight published the twelfth in a series of Brand Battles created by the team at Brandify. These studies pit two major national brands against each other to measure effectiveness across multiple channels of local marketing. I’d encourage you to check out the entire series and in particular the latest Brand Battle, where the two opponents are Starbucks and Dunkin’ Donuts.

As director of product at Brandify, I’m of course a little biased in my recommendation, but I think I can state objectively that Brand Battles do a good job of capturing two things: one, the real complexity of cross-platform digital marketing; two, the importance of a data-driven strategy in identifying meaningful objectives and tracking performance.

What I’d like to offer in this article is a nuts-and-bolts commentary that explains how Brand Battles are constructed and how each of their subject areas fits into the bigger picture of local marketing for national brands.

Topic One: Data Quality
In order to assess the quality of local data for a national brand like Starbucks, we first need to compile a list of all Starbucks locations across the United States, paying particular attention to the accuracy of our source data as regards NAPW information: business name, address, phone number, and local store web address.

We then compare the source data against what is published on major local sites including Google, Bing, Yelp, Facebook, and Foursquare. Though Brandify reporting capabilities extend to many additional sites, for purposes of Brand Battle analysis we focus on the sites that garner the highest consumer traffic. Every Starbucks listing on each site — more than sixty-five thousand listings in total — is matched against source data to determine whether store locations are listed accurately. We also gather data on whether the listing has been claimed on each site at the corporate level or by the store manager.

In the latest Brand Battle, we found that Dunkin’ Donuts had a greater level of completeness and accuracy in its online listings. An excerpt below shows a sample of the trouble areas for Starbucks, where on both Google and Bing more than 7% of locations across the U.S. could not be found, and among those we did find, a significant percentage had accuracy issues with addresses and phone numbers.


The significance of this aspect of the study is probably obvious. If customers can’t find you, they can’t do business at your store location. Perhaps surprisingly, Starbucks scored only 83 out of a possible 105 points in this section of the study; not an abysmal showing, but one that points to some needed improvements.

Topic Two: Local SEO
This part of the study dives into several areas, broadly covering the topics of onsite local SEO and search performance of owned properties. Metrics include the usability, content, accessibility, design, and load time of brand websites, store locators, and local landing pages.

We also look at performance in search on both Google and Bing for selected keywords, such as “coffee” and “breakfast” in the case of the latest Brand Battle, with geomodifiers such as city+state that allow us to measure how well a brand’s locations rank against other brands and local businesses in the same verticals.

As with data quality, search ranking analysis requires that we run thousands of search queries in order to capture all variations of keyword plus location across both search platforms. We aggregate all of that data in order to assign an overall score to each brand.

We place a heavy emphasis on search performance, knowing that for many consumers, the brand website is an afterthought. Especially when it comes to low consideration, frequent transactions like buying coffee, consumers are more interested in quick information about the nearest locations than they are in researching a brand’s website, so they often skip that step entirely. Still, it’s important to provide an optimal user experience for those consumers who do prefer to use the website and store locator.

Topic Three: Reviews
For the reviews segment of the study, we use APIs to check for review content on sites like Google, Yelp, and Facebook. One key metric is location rating, which is a star rating assigned by many sites to business listings consumers have reviewed. Similar to average rating, the location rating calculation generally includes those reviews the site determines to be trustworthy, and is sometimes only assigned once a location has amassed a certain minimum number of reviews. Location rating is the overall rating consumers will see when they view a business listing.

Below is a sample for Starbucks showing review metrics for multiple sites in the month of September 2016.


We also perform a comparative sentiment analysis to see how brands perform in consumer reviews for key terms relevant to the brand. In the current Brand Battle, those terms included “coffee” and “service.” We look at positive and negative review distribution to determine whether consumers associated positive or negative sentiment with each term.

The idea here is to use both quantitative and qualitative analysis to determine the impact of reviews as an aspect of local strategy. In the Brand Battles, we stop short at measurement, but as a standard practice we encourage brands to include review monitoring and response in their local marketing activities and to use the content of reviews to help create a feedback loop between consumers and store operations.

Topic Four: Advertising
For this topic, our analysis looks at both national and local advertising campaigns. If the brand does not have local ad spend, we look at national campaigns and estimated budgets, then assess which brand has a better opportunity to advertise locally based on budget, number of locations, and current data quality.

Local ads are proven to be much more effective than global or national campaigns at driving online customers to a brand’s physical locations. For this reason, advertising is a critical component in a holistic local marketing strategy.

Topic Five: Engagement
As with advertising, for social engagement we assess both national and local levels of the brand’s overall social campaign, focused largely on Facebook and Foursquare at the local level. Social metrics we examine include likes, check-ins, posts, and comments.

In the data quality section, we measured how effectively each brand had established a presence for all of its stores across the same social channels; here we are interested in determining whether brands have leveraged these social platforms to engage with consumers.

An important point to emphasize here is the need to ensure all local platforms are kept relevant by providing fresh content and using the tools available to communicate with consumers. A Facebook location page will not do much for a brand if it isn’t used as a channel for brand content and a forum for user engagement.

Damian Rollison is Director of Market Insights at SOCi.