How to tell what to fix first from your hotel's reviews

Why the small, cheap fixes come first

A refurbishment takes months to show up in reviews. A new routine, a check or a small purchase shows up in the next few weeks, and the complaints that keep coming back usually have fixes like that.

Two facts make it pay off. Booking.com says it weights reviews by date: the more recent a review, the bigger its impact on the overall score (Booking.com, how guest reviews work). And a study by Chris Anderson for Cornell's Center for Hospitality Research (2012) found that a hotel whose rating rose by one point on a 5-point scale could raise its prices by 11.2% and still keep the same occupancy or market share, using Travelocity data (Anderson, Cornell, 2012). That is an association measured on 2012 data, not a promise for every hotel.

So it helps to keep two lists: what can be done this week for little money, which goes first, and what needs real investment, which goes into the year's budget with the review data to back it up.

How to decide what to fix first

This method works in a spreadsheet or in a review management tool. What matters is the order of the steps, not the software.

1. Pull the last 90 days of reviews from the sites where you get the most: usually Google, TripAdvisor and Booking.com. If you have fewer than 50 reviews with text, stretch it to 6 or 12 months. Include complaints tucked inside good reviews ("loved it, but the wifi..."): they are often the easiest to fix.
2. Break each review into specific complaints. "Noise" cannot be fixed; "the air conditioning in the pool-facing rooms is loud at night" can. Note where and when it happens if the guest says: room, floor, area, shift or day.
3. Count reviews, not mentions. A complaint counts once per review even if the guest repeats it, and if the same person pastes their review on two sites, it counts once.
4. Pull out anything urgent. Food poisoning, a fall, pests, something broken that could hurt someone or a serious hygiene problem gets handled today, even if only one review mentions it. It stays out of the cost ranking.
5. Find the smallest fix for each complaint, with an owner, a cost and the time it takes. One fix can solve several complaints: "the shower runs cold in the morning" and "low water pressure" may be the same water heater.
6. Rank by cost and then by reviews solved. Cheap fixes that take days go first; among those, the one that solves the most reviews leads. Expensive work goes on a separate list, even if it is the most common complaint.
7. Pick 3 to 5 fixes for this week, each with an owner and a way to check it is done (a signed sheet, a supervisor's walk-round). More than five at once rarely sticks, and afterwards you cannot tell which one worked.
8. Check whether the complaint goes down. Compare the share of reviews that mention it in the weeks after the fix with the 12 weeks before, at 4 and at 8 weeks. If you get few reviews, wait longer before calling anything solved.

How to spot the complaints that keep coming back

A complaint is really recurring when it shows up in several different reviews, not when one angry guest makes it three times. To see that in a spreadsheet, you need one row per complaint and these columns:

With that sheet, three rules help separate real patterns from noise:

• A general topic counts with 3 or more reviews in 90 days. Fewer than that may just be a bad day.
• Something specific needs only 2 reviews, or 1 if checking takes minutes: "the shower in room 204" gets looked at today and closed.
• Look at the same month last year. If air conditioning comes up every August, the fix goes in before summer, not once the complaints arrive.

Read what guests praise too: what they love most is what you should not touch when cutting costs.

What the list looks like

Illustrative example, not taken from any real hotel. A 60-room hotel goes through its last 90 days of reviews and finds five recurring complaints:

The most common complaint, the bathrooms, comes last: it cannot be fixed this week. The first three cost next to nothing and together solve 20 reviews. And if one review mentioned a loose balcony railing, it would not go in the table at all: that gets fixed today.

Cheap fixes for common hotel complaints

These are starting points, not diagnoses: before spending anything, someone on the team checks where and why it happens.

Once a fix is done, replying to the reviews that asked for it and naming the specific change lets the next guest know. The guide on how to reply to your hotel's reviews explains how to do that without promising anything that has not happened yet.

What came out of a test report

This example comes from a PolarPuffin test report built on real data and anonymised: an all-inclusive beach hotel, with 204 reviews with text over 90 days from Google, TripAdvisor and Booking.com. These were the five fixes for one week:

None of them is a refurbishment: tasting the buffet before it opens, rotating the dishes, a five-minute talk each shift, an issue log and making sure there are two activities a day. All of them get done in the same week.

And this is how it gets measured (illustrative example): the complaint "hot water in the morning" came up in 26 reviews in the 12 weeks before the fix and in 3 reviews in the 8 weeks after. As the two windows are not the same length, compare them per week (about 2.2 reviews a week before and 0.4 after) or, better still, as a percentage of the reviews in each window. The rule for saying a fix works: complaints fall to half or less, and it is not chance.

With a spreadsheet or with a tool

If you get only a handful of reviews a month, a spreadsheet and an hour a week will do. Once reviews arrive in several languages and on several sites, counting by hand gets slow and it is easy to count the same one twice.

Many review management tools group complaints by topic and measure sentiment; the page on hotel review management software sets out what each one publishes, with its source. Grouping by topic covers steps 2 and 3 of this guide; ranking by cost and checking the result afterwards are usually left to the hotel.

PolarPuffin runs the steps in this guide every week: it reads Google, TripAdvisor and Booking.com reviews, suggests up to five fixes, each with an owner, a cost and the reviews behind it, flags anything urgent and checks at 4 and 8 weeks whether the complaint went down. It does not send surveys or do email marketing.

PolarPuffin costs $79 a month per hotel (Hotel plan), excluding VAT, with no minimum term, or $63 a month billed yearly, and you can try it for 30 days with no card. More on how it works and pricing.

Common mistakes when prioritising guest complaints

• Starting with the most common complaint without looking at cost. The most mentioned issue is often the most expensive; while it gets decided, the cheap complaints keep coming.
• Spending before checking the cause. "Bad night's sleep" can be the mattress, the light, the door or the air conditioning.
• Changing ten things at once. Afterwards nobody knows what worked.
• Measuring the number of complaints instead of the share. In low season there are fewer complaints because there are fewer reviews.
• Replying with a promise of a fix that has not happened. The next guest checks and says so.

For each site's score, see the guides on your Booking.com review score and your Google and TripAdvisor rating.