How NLP-based VoC programs can help hospitality brands meet customer expectations during COVID-19


Share on LinkedIn

COVID-19 has had an extensive impact on the hospitality industry. Shelter-in-place restrictions, along with event cancellations, reductions in business travel and widespread job losses have altered how and why people travel. Faced with lower occupancy, shifting safety measures, and catering to an increasingly local clientele, hotels and hotel-adjacent businesses are having to rethink their business models.

Since 89% of businesses compete mainly on customer experience, it’s vital that hospitality providers deliver on guest expectations. A solid Voice of the Customer program (VoC) underpinned by AI and Natural Language Processing (NLP) can help hospitality businesses do just that. Below is a roadmap for implementing an effective VoC program for the COVID-19 era.

1. Identify who’s responsible for handling guest feedback

First, identify who is responsible for managing customer feedback within your organization. This responsibility may fall upon or be shared among your sales, marketing, operations or customer service team. Importantly, be sure to understand what kind of feedback each team actively seeks out, and to what end. For example, your marketing team is most concerned with brand perception, your sales team with driving conversions, and your operations team with actionable items. Be mindful of potential silos or biases and try to get your teams working together toward a common goal.

2. Select a specific question to answer

Next, identify a pressing question or issue to examine. Choose something specific and actionable – something that will help improve your products and services. These might be something like:

  • What are the top concerns and expectations of my customers during COVID-19?
  • What changes should we make to our hotel rooms or spa in the next quarter?

Avoid being too broad or non-specific, as you won’t receive the targeted feedback you need to make informed changes.

3. Gather and prepare your data

There are dozens of potential sources from which data can be gathered, but not all will be ideal for answering your question. Being selective about your data sources helps you avoid gathering irrelevant and ineffectual data. For example, if your question relates to brand reputation, Twitter is a great data source. If your theme is customer satisfaction, look at survey responses or online reviews. Chat conversations, call transcriptions, market research and complaints are also viable sources – but it depends on what you’re looking for.

4. Choose the tools for gathering and preparing your data

There are myriad data analysis tools out there, and what works for your organization will depend on the size of your business, the scope of your question, your available resources and whether you’re conducting your analysis in-house or outsourcing it. Some popular in-house tools include SurveyMonkey or TypeForm for survey analysis, SocialGist and Sprinklr for social media analysis, along with various NLP analytics, visualization and reporting tools. Platform solutions such as MyHotel and Revinate offer an all-in-one approach, although often without the customization of NLP solutions. Depending on the project, you might also want to outsource everything to a third-party provider.

5. Prepare your data and configure your analysis

Taking the time to properly select, check, clean and standardize your data is a critical first step that will both save you a lot of time and ensure that your results are accurate. This is particularly true for AI and NLP-based tools, which work best when applied to large amounts of data.

Common data problems include:

  • Content using the wrong language
  • Using a mix of formats and styles
  • Inadvertently including advertisements
  • Inclusion of irrelevant data due to non-specific keywords (eg “Paris Hilton” showing up in Hilton Hotel data)
  • Spam content
  • Missing metadata (e.g. location, date, room number and star rating).

Additionally, sophisticated tools such as AI and NLP-based solutions will allow you to configure or “tune” your analysis so that the results you get are properly tailored to your question. This may include adding staff names or menu items, or creating unique COVID-19 specific queries to quickly identify relevant themes.

6. Process and analyze your data to identify themes

Once your data has been processed, examine the results to identify and examine recurrent themes and topics, as well as the corresponding sentiment in relation to each of these. For example, if bathroom cleanliness, social distancing, and general manager repeatedly show up, is this in a positive or negative way? Analyze whether both sentiment and frequency has changed over time, if anything new has emerged, and what the current sentiment is.

Importantly, while basic tools can give you an overall picture of how your organization is performing, AI and NLP-based solutions allow you to take a deeper dive to identify and focus on moments of negative sentiment hidden in otherwise positive reviews – things that might get overlooked if you were simply looking at star ratings, NPS scores, or other high-level trends.

7. Draw conclusions and create a strategic response

Once you’ve isolated key themes, you can start making strategic decisions about what needs to be fixed immediately, what can be monitored over time, and what is simply unfeasible or off-brand. In doing so, be sure to consider your data over time, and within the larger context of your brand, business, positioning, and the broader competitive landscape. For example, a surge in complaints about bathrooms or value for money may indicate a shift in market expectations.

Now that you’re aware of a potential problem, you can consult with your decision-makers to come up with a method, timeline and budget for addressing the issue.

VoC helps you protect your brand from COVID-19

For hospitality businesses, VoC tools offer an important window into the guest experience and can help brands see where their offering meets or exceeds expectations, falls short, or could adjust to meet a gap in the market. Given the enormous changes the industry has experienced during COVID-19, and the rapidly shifting norms, expectations and standards today’s hospitality brands are subjected to, an ongoing investment in VoC can be invaluable in identifying areas for improvement.

Andrea Kulkarni
Andrea Kulkarni draws upon 20 years of diverse experience in designing, developing and marketing business and education software. At Lexalytics she leads a variety of initiatives, spanning product and market positioning, documentation, web analytics, growth marketing, and CRM operations. Andrea holds a degree in Learning Design and Technology from Stanford University.


Please use comments to add value to the discussion. Maximum one link to an educational blog post or article. We will NOT PUBLISH brief comments like "good post," comments that mainly promote links, or comments with links to companies, products, or services.

Please enter your comment!
Please enter your name here