Homeowners mostly hire contractors through a word of mouth. We look for these references because of the spotty reputation the renovation industry has with users. There is another way to find the right home renovation professional for your project.
Then came reviews! And, wow, did they change the behavior, though that’s hard to prove. In a survey of home renovators conducted by Kukun, 91.2% said they would not feel comfortable hiring someone just through a review. They prefer to blend the two approaches.
However, professionals are becoming more tech savvy. They have profiles across multiple platforms such as Google+, Facebook, Yelp, the BBB, and many other sites. Professionals actively encourage homeowners to post reviews for them – especially good ones. This way, the positive reviews can bring them more customers.
Due to the spread of reviews across multiple websites, the average homeowner has to go to several different places (and most are not aware of all of them). Going through all the reviews to find what the reviews are essentially saying can be dizzying and time-consuming.
The Problem = Motivation
Motivation is a problem because reviews are often ambiguous, and going through them can further confuse readers. Reviews are also jampacked with irrelevant and flowery comments with a lot of irrelevant and unnecessary information. As a result, Kukun decided to take on the effort of summarizing the results to the homeowner using what is typically referred to in the Data Science world as “Sentiment Analysis”.
Solution: Sentiment Analysis
Our algorithm and approach looks for key opinionated snippets within these reviews and presents them to the user in a more intuitive way. For instance, a snippet could be about homeowners talking about the quality of service provided by the professional, or it could be about whether the contractor adhered to the time-frame and budget. However, when there are large number of reviews, there can be lots of worthwhile snippets. In order to distinguish these snippets, we rank them based on the “usefulness” of the sentence. We give a high rank to those snippets that occur in the most reviews and are informative.
Apart from this, we also provide a final verdict on four key attributes, namely price, quality of service, on-time completion, and aesthetics. For every opinionated snippet, we find out whether it is positive, negative, or neutral using a machine learning approach. Based on number of positive, negative, and neutral reviews occurring in each category, we give a category score. With this, any user is able to understand the general opinion of other homeowners about the professional’s on-time completion, budget adherence, etc.
We believe that by using our tool homeowners can make an informed decision at one place rather than trying to juggle a number of websites. Any user can simply visit a professional’s profile at Kukun and find an aggregate review along with attribute scores. In addition to this, we have plenty of other helpful information for homeowners about the professional, including photos, licenses, permits, etc. So you can hire the right pro at the right price.
Along with Sentiment Analysis, looking at the permit history of every professional can expand your search network and gives you direct access to real references with real homes so you no longer have to depend on your limited network.