Case Study – Analyzing Customer Reviews for a Restaurant

Introduction

This is the first in a series of “Case Study” posts where we get hands-on to see what we can learn from public Customer Review data.   Customer Reviews are important to a business.   The last time you did anything that involved interacting with a business what did you do first?  You most likely checked your favorite search or social app for recommendations or at least for social proof that the business you are going to interact with has a high social approval rating.   Businesses who ignore their bad approval rating are way less likely to attract new and returning customers.

Approach

 We are going to use the following approach:

  • Select a business category.  We will select a Restaurant to analyze.  
  • Identify a business that has 500+ customer reviews for their Google business page
  • Use the okaxis customer insights platform as our tool of choice to automatically gather the required data and generate insights for us.  Mainly because we don’t have a team of data engineers and data scientists just sitting around 🙂

Goals

  • Gather basic Voice of the Customer (VOC) summary statistics regarding the customer review data we are analyzing.  How many customers are interacting with this business?  How often do their customers interact with them?  
  • Understand the overall customer sentiment.  Do reviewers generally like or dislike this business based on the reviews?
  • Understand what specifically reviewers dislike.
  • See if we can discover any interesting trends in the customer reviews.
  • And most important, determine what actionable insights can be derived from the customer reviews which may help the business improve.

Voice of the Customer (VOC) Summary Statistics

 Ok let’s get into this.  

We selected a Restaurant that has 858 customer reviews so that meets our goal of at least 500 reviews.   Think about this for a second.  That’s 858 direct, verbatim samples of customer feedback on their business.   What a gift.  

We can also see that 855 are unique customers so that means a few customers felt compelled to share more than one review.

The first thing to notice here is this business has had steady growth in terms of monthly customer review count. For a business to get an average of 17 reviews a month it clearly indicates they are remarkable to their customers.  Well done.

So far this month, they received 11 reviews. Last month they had a total of 24 reviews which is over their average.   As expected, we see only one review source- 100% of the reviews came from Google.    It would be interesting to connect additional data sources (like Facebook for example) to see if we see the same steady growth.  Okaxis makes it super easy to connect to other data sources.  Simply click a button and connect your account for that data source and poof the reviews are now in okaxis.

okaxis voice of the customer (voc) summary statistics

Another thing to understand is what does this look like year over year?  An average of 17 reviews a month would net about 204 reviews per year.  Looking at the graph below, we can see that reviews for this business started ramping up in 2016 and continued to grow thru 2019.   Then in 2020 saw their first decline in the count of customer reviews.  In fact, looks like they only received just over half of the reviews they received from the year before.   This is a similar trend we see to other restaurants and businesses in general due to the Covid 19 pandemic. 

This will be an interesting thing to look at when we dig into the data further.

okaxis - voice of the customer (voc) YOY trend

Customer Sentiment Summary

Sentiment is a measure of how ‘Positive’, ‘Neutral’, or ‘Negative’ your customers feel about your business or brand.

It’s great to see a business with an overall 80% positive sentiment (actual is 79.8%).  The last 6 month average has dropped by .7 points (from 79.8 to 79.1) which is not significant but something to continue to watch.

We can see an overall 12.8% Neutral Sentiment Average which has increased to 14.8% over the last 6 months average.  And finally, an overall 7.3 % Negative Sentiment Average which has decreased to 6.1% over the last 6 months average.

okaxis - overall sentiment summary

Looking at the 12 month trend for negative sentiment, we can see a few spikes.  We will note these spikes for now and dig into the details as we progress to see if we can find anything actionable which would help this business improve.

okaxis - overall sentiment trend graph

Customer Sentiment – Subjectivity

Subjectivity is a score between 0 and 100% where 0 is very objective or factual and 100% is very subjective or opinion. 

We can see an overall 61.6% Total Subjectivity % Average which has increased to 63.1% over the last 6 months average. For this type of business and for customer reviews in general, a subjectivity score of 60% is on benchmark.   The Subjectivity average for 12/20 is lower where 1/21 is higher than other monthly averages so we will take a look at this further to see what is happening.

okaxis - subjectivity and word frequencies

Customer Sentiment – Word Frequencies

Looking at positive and negative word frequencies is a great way to quickly identify sentiment trends.

 Top Positive Word Frequencies – the top 10 positive words in this businesses reviews are:

  1. Great
  2. Good
  3. Nice
  4. Excellent
  5. Delicious
  6. Friendly
  7. Amazing
  8. Awesome
  9. Love

Scanning other positive high-frequency words (words that occur in multiple reviews) that we may want to drill into would be:  Fantastic, more, better, fun, beautiful, alternative, favorite, huge, outstanding, perfect, clean, fresh, special, unique, sweet…

Top Negative Word Frequencies – the top 10 negative words in this businesses reviews are:

  1. Little
  2. Bad
  3. Slow
  4. Cold
  5. Disappointed
  6. Few
  7. Chicken
  8. Terrible
  9. Expensive
  10. Poor

Scanning other negative high-frequency words (words that occur in multiple reviews) that we may want to drill into would be:  Average, small, unfortunately, rude, wrong, extremely, empty, worse, expected, disappointing, missing, boring, horrible, shocked, sour…

Let’s dig into some of these high-frequency negative words to see if we can find any trends or actionable improvement areas.

okaxis lets you easily search across all your reviews…  This can save you a ton of time especially when you have thousands of reviews from multiple data sources.

okaxis - customer reviews search for little

Searching the keyword “little” filters the reviews down to only those that contain that word:

  • “Just one little annoyance. Sour cream came with my meat loaf tacos and I asked for salsa instead and was charged 50cents!”
  • “As far as my sweet potato fries they could have cooked them a little bit longer because they were soggy.”
  • “The baked tomato soup was delicious and the broccoli was good also a little small though.”
  • “My sandwich was good but a little pricey for a sandwich and chips.”
  • “I’m on a fixed income and it’s a little out of my range, but worth the occasional splurge.”
  • “…green beans just a little too crunchy for me”
  • “Steer clear of most pastas, very pricey with very little substance.”
  • “Great variety. Expedient service. A little pricey for quality. Outdoor seating relaxed environment.”
  • “Food was okay. I expect a little more for the price.”
  • “Good food, nice menu selection, can get a little loud”
  • “good food, good service, good atmosphere, but a little on the pricey side for what it is!”
  • “A little pricey , but worth it .”
  • “little bit pricey.”
  • “A little pricey, food good”
  • “Bar area can get a little loud.”
  • “Prices are a little high”

 It is easy to quickly note some trends in negativity here regarding:  price, portion size, atmosphere

okaxis - customer reviews search for slow

Search keyword “slow”:

  • “Great food. Service was slow. Prices are a bit high.”
  • “Nice place. Good food. Service is slow to seat customers.”
  • “the waitress was nice but extremely slow…”
  • “Food was good I had the meatloaf. Service was slow…”
  • “…service was slow, but overall a great place to have a meal.”
  • “Slow service & food just fair.”
  • “Little slow on the service, they could ramp that up but the food and the quality is there.”
  • “Service was slow and they weren’t under staffed even though it was busy.”
  • “Service was a little slow, but overall a good place to eat.”
  • “Hostesses were very slow to fill the 4 open tables.”
  • “Super slow service. Waited 15 minutes for anyone to even acknowledge us..”
  • “Service was a little slow, but it was Friday night.”

 Again, we can quickly note some trends in negativity here regarding:  primarily Service…

okaxis - customer reviews search for cold

Search keyword “cold”:

  • “Food was served cold”
  • “Ribs dry over done. Chicken potato cold so so service. Cold. Most people eating with their coats on. Did not live up to our expectations”
  •  “…they seated us at a table that was extremely cold”
  • “…the table was near the door and it was cold.”
  • “…The entree was not only lousy but cold.”
  • “…My mashed potatoes were also cold.”

  In this case, we note some trends in negativity here regarding:  Food preparation and atmosphere…

Search keyword “poor”:

  • “Very poor food, don’t get fish and chips. Terrible, three fish sticks and a hand full of French fries.”
  • “all around it was a poor experience and have had poor experiences the past few times I’ve dined there.”
  • “Food Fair … served sediment in wine….poor!!!”
  •  “Good beer list. Poor wine list.”

And finally, we note some trends in negativity here regarding:  Food, Wine, Experience.

Visually Discovering Trends

 “Word Clouds” help you to visually understand word or phrase frequency in your data… The more times a word or phrase is mentioned, the larger the text.  

Let’s look at a few out of the box Categories that okaxis provides.   Okaxis knows this is a Restaurant so it gives us a “Menu Item Mentions” category.   We can see that there were 584 mentions pertaining to menu items.   Additionally 99% of the discussion around these terms were positive. 

The top terms (largest) talked about on the menu are beer, drinks, crab, steak, chicken, sandwich, and salad.   Okaxis allows you to customize categories and also create new ones.   We could add our specific menu items to get a more in depth understanding.   Note that we can also see the details under the word cloud so we can see the exact customer reviews.  We can search this list or page through all results.

okaxis - customer reviews trends - menu

Looking at the “Reputation Promotors” category, which shows us positive opinions related to company reputation, we see that there were 37 mentions pertaining to this category. 

The top terms (largest) talked about regarding company reputation are highly recommend, would recommend, definitely recommend, top notch, very polite, very knowledgeable, etc

okaxis - customer reviews trends - reputation

Looking at the “Customer Service Promoters” category, which shows us positive opinions related to customer service feedback, we see that there were 143 mentions, of which 99.3% were positive, pertaining to this category. 

The top terms (largest) talked about regarding customer service feedback are friendly, great service, good service, excellent service, very pleasant, courteous, etc

okaxis - customer reviews trends - service

 

Looking at the “most common phrases in low rated” category, which shows us opinions mentioned in low rated feedback (reviews with a 1 or 2 star rating), we see that there were 39 mentions.   The top terms (largest) talked about in low rated feedback that we may want to explore further are Crab legs, food quality, service, etc.

okaxis - customer reviews trends - low ratings

 

Lets dig into one of these to get a better understanding…  Why did Crab legs make the list?   Looking at the reviews, it looks like the restaurant has a Tuesday night all you can eat crab leg special.   Apparently there is a lot of passion around the expectation of getting something only to find out you cannot have it because they sold out among other concerns…  🙂

  • “ran out of crab legs”
  • “We both got crab legs & they were very very salty.”
  • “Don’t advertise all you can eat crab legs on a Tuesday and be out of them at 6:30pm…”
  • “The worst service i have ever had. Crab leg night and had to wait 25 minutes for second plate.  They were not that busy. Never given an apology others in the party were never asked about more drinks or dessert while the table was waiting. VERY DISAPPOINTED!!!”
  • “crab legs were small, dry and overdone”

Key Findings & Action Plan

Ok so how did we do?   I’d have to say, looking at the goal we set out to achieve, we nailed it.

We were able to:

  • Gather basic Voice of the Customer (VOC) summary statistics on this business we analyzed. We now know how many customers are interacting with this business AND how often their customers interact with them.
  • Understand the overall customer sentiment for this business.  It is very clear that reviewers really  like this business based on their reviews.
  • Understand what specifically reviewers dislike.  This business had very few negative reviews overall.   When reviewers did complain they were focused in these areas: price, portion size, atmosphere, food/food prep, wine list, and service…
  • Discover some interesting trends in the customer reviews.  We now know that the majority of  reviewers indicated that this business was friendly, had great/excellent service, was very pleasant and courteous.   Additionally reviewers would highly recommend them, found them to be top notch, very polite, very knowledgeable, etc
  • Focus in on many areas where the business could double down on what it is going good and also improve where reviewers high-lighted short comings.   Crab legs anyone?

Leave a Comment

Your email address will not be published. Required fields are marked *