Analyzing McDonald’s vs Burger King with Symanto

We’re big believers in specialization at Datashake, and while we focus on data aggregation and packaging, we rely on trusted partners to achieve the data analysis component. Symanto is one such partner and we’re going to do a deep-dive into their capabilities in this blog post. They are a leading industry expert in Deep Learning, Natural Language Processing & Psychology at the forefront of AI development since 2010.

While we focus on the data extraction layer, Symanto’s focus is on generating insights from said data:

Symanto has been using Datashake for a couple of years and in their own words, this is how they benefit from our partnership:

1. Scalability: Datashake is designed to handle large volumes of data quickly and efficiently, which is very beneficial for our clients who are short in time and need to quickly collect review data and analyse it.
2. Cost savings: By partnering with Datashake, we wanted to make sure we offer a 360 solution to our clients, by allowing them to not only analyse but also collect data at a low cost.
3. Customisation: The Datashake crawling requirements allow for adding specific filters when collecting the data which can be extra helpful for running the analysis afterwards and getting more granular insights.
4. Access to over 100 data sources: Datashake gives access to new sources of data to our customers that they might not have been able to analyse before.

To make sure this analysis hits home, we’ve chosen a real world use case analyzing the reviews for 3 McDonald’s locations vs 3 Burger King locations. I’ve included the raw data below in case you want to take a look or analyze it yourself.

These are the steps we’ve taken to complete this project:

1. Choose the locations

We chose the following locations for McDonald’s and Burger King, based on 3 cities:

1. New York (0.2 mile radius)
   • McDonalds: 688 8th Ave, New York, NY 10036, United States
   • Burger King: 327 W 42nd St, New York, NY 10036, United States
2. London (0.1 mile radius)
   • McDonalds: 48 Leicester Square, London WC2H 7LU, United Kingdom
   • Burger King: 17-21 Leicester Square, London WC2H 7LE, United Kingdom
3. San Francisco (0.5 mile radius)
   • McDonalds: 609 Market St, San Francisco, CA 94105, USA
   • Burger King: 35 Powell St, San Francisco, CA 94102, United States

2. Data collection

We use our Review Index API to collect the required data – all it takes is 1 API call per location to collect the reviews across sites like Google, Facebook, Yelp, Tripadvisor and more. Here is a sample API call (decoded for sake of clarity):

curl 'https://reviewindexapi.datashake.com/reviews/?
api_key=x&
callback=https://eo7z26l26rzn456.m.pipedream.net&
name=McDonald's&
address=609 Market St, San Francisco, CA 94105, USA'

Attached are the outputs for these requests grouped by city:

It’s worth noting that our Review Index API can be used both to fetch data on a specific address (as shown above), as well as all addresses for a given business. This means that you can supply just a company name like McDonald’s and we’ll return the reviews for all of their 30k+ locations.

Data analysis

On to the part where Symanto shines! Once the data is ingested into their platform, they can help answer questions like:

1. What do people talk about?
2. How do people feel about a brand?
3. What are the emotional and rational drivers of purchasing?
4. Are consumers recommending a brand to others?
5. Are people emotionally connected to a brand?
6. What is the rating impact?

Below is a snapshot of their dashboard so you can see the extent of analysis that can be performed, and we’ll go through some examples below.

Diving into the customer experience, we can see that Product is by far the most important aspect for both brands at 44% for Burger King and 30% for McDonald’s. Within that grouping, customers compare McDonald’s and Burger King in terms of food quality, with some preferring the quality at McDonald’s; they also compare in terms of Speed of service:

Next, we use a sentiment detection that goes beyond keyword sentiment and rather looks at sentences and paragraphs as a whole, to truly understand what sentiment is associated with what topic. Burger King has the best net sentiment and highest share of positive mentions:

We can also understand how consumers describe their emotional experience with a product or brand: in our current project, in 49% of the conversations people express negative emotions (mainly sadness) when talking about McDonald’s:

This is exemplified by a review from Mathieu at the McDonald’s San Francisco location which stated “Sadest interior I have seen in a long time …” and a review from Sam at the McDonald’s London location stating “Sadly it’s still far smaller than you would expect of Leicester Square, so a short queue is common any given evening and seating inside is limited. On this occasion the Burger King on the other side of the square is far bigger and more comfortable, though only when they have the upstairs open.”

Based on psycholinguistic consumer modules, Symanto uses further deep-learning clustering to perform consumer segmentation. 22% of the Burger King consumers are at-risk – frustrated customers who are emotionally disconnected and tend to share their frustration rather than opinion, which can be highly influential for other emotional customers. The topics which influence their opinion the most are Food, General Service and Staff Friendliness. McDonald’s has even higher number at 25%, with people further referring to the Ordering process using terms as slow, messy, poor and long waiting.

By correlating the average rating with the topics mentioned, we can start to understand the impact different topics have on your company, competitiveness, and customer base. The average rating can identify areas where you are doing well and areas that need improvement (customers are dissatisfied with a particular aspect of the business):

Finally, Symanto’s most recent development – the insights assistant. By processing your custom dataset, the insights assistant provides immediate answers to your questions, allowing you to explore unstructured data intuitively. 

Conclusion

I hope this has been an informative blog post which shows a real world use case of how powerful analyzing data that is collected through Datashake in our partner Symanto can be. As a next step, you can sign up for Datashake to try our data collection capabilities, and Symanto to try their data analysis capabilities.