25
Aug 2026
How OASIS Extracted a Nation’s Worth of Accommodation Review Data in Under One Week with Mentionlytics
How OASIS used Mentionlytics to extract 160,000+ accommodation reviews in under a week.
CASE STUDY
How OASIS Extracted a Nation’s Worth of Accommodation Review Data in Under One Week with Mentionlytics
Client: OASIS Optimisation
Contact Person: Pierre Coetzee, Consulting Partner
160,000+
mentions/reviews extracted in under 1 week .
5 weeks
from setup to proof-of-concept for a national tourism ministry.
One method.
A scalable review intelligence methodology, ready for any industry.
Summary
- OASIS Optimisation, a South Africa-based research and analytics consultancy, was tasked by a tourism ministry to create a proof of concept that mapped visitor experience across the country’s entire accommodation sector, spanning hundreds of hotels and resorts.
- Traditional research options, whether custom web scrapers with limited database access or expensive primary research involving instrument design and fieldwork, were either too unreliable or too costly to make the project viable.
- By integrating Mentionlytics, OASIS extracted over 160,000 mentions/reviews, spanning review platforms and social media, across the target country’s hospitality market in under one week, at a scale that would have taken months through conventional methods.
- The project delivered a complete proof of concept for the tourism ministry within 5 weeks, benchmarking visitor experience across hotels and resorts against competitor destinations, with custom topic modeling built on top of the Mentionlytics data export.
- OASIS now has a standardized, repeatable workflow for ratings and review intelligence that can be adapted to new industries and clients, with Mentionlytics as the foundation for rapid, reliable, and cost-effective data extraction at every stage.
About OASIS Optimisation
OASIS Optimisation is a South Africa-based research and analytics consultancy that helps organizations make better decisions through data. Their work sits at the intersection of strategy, consumer insight, and quantitative research, with a strong focus on hospitality, tourism, and destination intelligence.
When clients need to understand what people actually think and feel, OASIS builds the frameworks to find out.
The Challenge: Find and Extract Thousands of Reviews in Different Languages at a Regional Level
Pierre Coetzee, Consulting Partner at OASIS, and his team were approached by a tourism ministry with an ambitious request: understand the full visitor journey, starting with accommodation.
The goal was to analyze how travelers were experiencing hotels and resorts across the country, and then benchmark that against other destinations.
The project needed scale. We’re talking about hundreds of properties, thousands of reviews, multiple languages, and data that had to be aggregated at a regional level to be meaningful.
Without turning to a review monitoring tool, OASIS had two realistic options, but neither was the best fit:
- Build a custom web scraper. However, the problem was that scrapers couldn’t access every review database effectively. They demand constant upkeep, and the data quality is never guaranteed.
“You’d have to put a lot more time and effort into making sure it collects the right data.” — Pierre Coetzee, Consulting Partner at OASIS Optimisation
- Take the traditional primary research route: design a measurement instrument, field the research, and then run the analysis. That’s the gold standard for precision, but it comes with a price tag to match.
“It’s definitely a lot more time-consuming and expensive. You’d have to design your own measurement instrument, then collect the data through a primary research methodology, and then analyze the data. It’s definitely a lot more expensive to go that route.” — Pierre Coetzee, Consulting Partner at OASIS Optimisation
What OASIS needed was a way to get reliable, structured review data at scale, quickly, and without the overhead of building or fielding their own research from scratch.
The Goal: Create a Proof-of-Concept within a Realistic Timeline Without Sacrificing Data Quality
Two things mattered most on this OASIS project:
- Speed and cost-efficiency. The proof-of-concept had to be delivered within a realistic timeline and budget, which ruled out traditional primary research entirely.
- Data quality at scale. The team needed review-level data (ratings, raw text, property identifiers, date ranges) that could be rolled up to a regional view and sliced by property when needed.
In addition to this project, Pierre and his team wanted to build a methodology that could serve their future clients, so the team won’t have to reinvent the wheel every time they are tasked with similar projects.
Why Mentionlytics
OASIS evaluated their options and chose Mentionlytics as the most practical path forward for this kind of ratings and review extraction at scale.
Unlike general-purpose social listening tools, Mentionlytics offered something the alternatives couldn’t: access to structured review data on accommodations across the country they targeted, delivered quickly, without requiring OASIS to build and maintain their own data pipeline.
The fit became clear almost immediately once they were set up. The keyword logic allowed them to pull data at the level of granularity they needed, and the review-level exports gave them exactly the raw material required to run their own modeling on top.
"Pretty much everything that was in there added value. We needed the rating, the date ranges to analyze trends over time, the raw text for classification, and then a way in which keywords could be used to roll things up at an aggregated level."
Pierre CoetzeeConsulting Partner at OASIS Optimisation
The Solution with Mentionlytics
Rapid Data Extraction at Scale
The first phase of the project was all about getting the right data, fast. OASIS used Mentionlytics to extract reviews about accommodations in the target country, with each record including:
- The review ID
- Numerical rating
- Raw review text
- Property identifiers
Keywords were configured to allow aggregation at the regional level, which was the primary unit of analysis for this project. That meant OASIS could roll up the data across an entire region or drill down to individual hotel performance, should the client need it.
From setup to having clean, usable data in hand: less than one week!
In the screenshot below, you may see the total volume of reviews the OASIS team was able to collect in less than a week (historical and latest data combined).
That timeline alone changed the economics of the project. No fieldwork, no custom infrastructure, no waiting.
Platform Validation Before Deep Analysis
Before OASIS moved the data into their own modeling environment, they used the Mentionlytics platform to do a first-pass sanity check. They explored keyword combinations, validated that the data being pulled matched the brief, and got an early feel for the themes emerging in the reviews.
By exploring different keyword combinations and zooming out of review platforms only, Mentionlytics provided OASIS with rich mention data from social media users who shared their experiences on YouTube, Instagram, X (Twitter), and other popular social channels.
This step was of utmost importance because catching keyword gaps or data quality issues before running topic models saves weeks of rework downstream.
The total number of reviews (including social media mentions) that OASIS gathered through Mentionlytics was more than 160K.
Especially, those reviews posted on social media wouldn’t be accessible to OASIS through custom web scrapers or other traditional research methods.
Custom Topic Modeling
Once the raw data was validated, OASIS imported it into their internal modeling system and got to work.
Every review was classified into specific topics. Ratings and sentiment were then analyzed against those topics to identify where guests were satisfied, where they were frustrated, and how the target country compared to benchmark destinations.
The Results: Proof of Concept Delivered & a Methodology Built to Scale
Total time for modeling, translation, and dashboard creation: approximately 4 weeks.
Total project time from setup to first deliverable: 5 weeks.
The tourism ministry received a decision-ready view of how visitors experienced accommodation across the country, broken down by region and topic, with sentiment benchmarked against competitor destinations.
But the outcome that OASIS values most is what the project left behind internally.
"It’s a standardized process now from our perspective. We have a robust process, and the entire process will go a lot quicker now because we've now set up the models and the scripts."
Pierre CoetzeeConsulting Partner at OASIS Optimisation
They now have a proven, repeatable workflow for extracting and modeling ratings and review data at scale. The models are built. The scripts are ready. Future projects in different industries or for different clients don’t start from square one; they start from a working playbook.
One practical improvement they are already planning for multilingual projects: requesting both the original language and the translated version from Mentionlytics upfront, rather than handling translation downstream.
A small process change that will save time on every future project that involves non-English reviews.
Their Advice
If you are considering Mentionlytics for large-scale data projects, Pierre’s takeaway is simple: the platform is the fastest way to get structured review data at scale without the cost and lengthy process of primary research.
"We can now take this proven concept, tweak it for different industries, and roll it out. We can now use a defined process and find other clients that can see value in it."
Pierre CoetzeeConsulting Partner at OASIS Optimisation
Ready to unlock the true potential of social media for your brand? See what Mentionlytics can do for you.
14-days trial No credit card required Cancel anytime