Hi, thanks for your interest in how we calculate an experience's ranking score. It's at the core of Rankers so I'm pleased you're curious.
The ranking score percentage is used to compare and sort experiences in ranking tables. It is not necessarily a direct measurement of the quality of a particular experience as rated by its customers. I've found it a useful tool to allow me to find the best experiences with confidence. But I've also found it important to read the customer reviews before making any final judgements!
We calculate an experience's ranking score using a multi-factor data model instead of a raw data average (mean). This model takes into account several important questions. For instance - is there a trusted body of reviews? What is the age of a review and is the review from a credible source?
Below you'll find details around some of the important factors that went into calculating the ranking score for Haast.
If you have any questions or comments about our ranking score calculation please get in touch at info@rankers.co.nz. We don't believe this is perfect or complete so we're always interested in ways we might make improvements.
Haast
Valid Reviews
20 Valid Reviews
The Haast experience has a total of 20 valid reviews. There are no invalid reviews that are excluded from the calculation. Reviews can be excluded only when a reviewer is not verified or after an investigation by our team determines the reviewer is not genuine.
Below is the distribution of ratings for the 20 valid reviews:
| Rating | Count | Percentage | |
|---|---|---|---|
| 10/10 | 5 |
|
25% |
| 9/10 | 4 |
|
20% |
| 8/10 | 3 |
|
15% |
| 7/10 | 4 |
|
20% |
| 6/10 | 0 |
|
0% |
| 5/10 | 3 |
|
15% |
| 4/10 | 0 |
|
0% |
| 3/10 | 1 |
|
5% |
| 2/10 | 0 |
|
0% |
| 1/10 | 0 |
|
0% |
78.00% Average
The raw data average (mean) for all the Haast valid reviews is 78.00% and is based on 20 valid reviews. This value is not used to calculate the ranking score and it only provided here as a comparison to the weighted average.
Face-to-Face Reviews
18 Face-to-Face Reviews
The Rankers team meets with travellers while they’re in New Zealand and conducts face-to-face surveys. These reviews, in our opinion, are the most trusted in the industry and represent a critical control sample. To our knowledge, we are the only travel review website in the world that has gone to this extent.
More about face-to-face reviews
Within the 20 valid reviews, the experience has 18 face-to-face reviews collected during interviews by our team.
Below is the distribution of ratings for the 18 face-to-face reviews:
| Rating | Count | Percentage | |
|---|---|---|---|
| 10/10 | 3 |
|
17% |
| 9/10 | 4 |
|
22% |
| 8/10 | 3 |
|
17% |
| 7/10 | 4 |
|
22% |
| 6/10 | 0 |
|
0% |
| 5/10 | 3 |
|
17% |
| 4/10 | 0 |
|
0% |
| 3/10 | 1 |
|
6% |
| 2/10 | 0 |
|
0% |
| 1/10 | 0 |
|
0% |
75.56% Average
The raw data average (mean) for all the Haast face-to-face reviews is 75.56% and is based on 18 face-to-face reviews. This value is not used to calculate the ranking score and it only provided here for comparison purposes.
Weighted Average
83.54%
Rankers calculates a weighted mean as a base average on which we can improve. Individual review's ratings are given a weight based on several factors. The weight of a review determines the overall impact it'll have on the final weighted average.
Recent reviews have more weight as they are more relevant and reflect the experience as it currently operates. Over time reviews become less relevant and loose their impact on the ranking score.
Low rating reviews carry slightly less weight. This dampens the effect of very low ratings for every experience across the board. This is especially important when the experience has few reviews overall and a single negative rating can grossly mischaracterise an experience. Consistent poor reviews will still result in the experience receiving a comparitively low ranking score.
Credible sources provide reviews that can be trusted. If we have verified a reviewer is genuine via a face-to-face meeting then the review carries additional weight.
| Reviewer | Rating | Age | Relative Weight |
|---|---|---|---|
| Veronica | 10/10 | 2271 days | 100% |
| Markus Johannes | 8/10 | 3567 days | 37% |
| Anna Grafenstein | 5/10 | 3911 days | 16% |
| jefft | 7/10 | 3916 days | 19% |
| Henrik Bours | 7/10 | 3930 days | 19% |
| Annika Marz | 5/10 | 3938 days | 15% |
| Andrew Magill | 10/10 | 3952 days | 19% |
| Glyn and Karen Farlow | 3/10 | 3962 days | 10% |
| Kiernan Matthew | 5/10 | 3973 days | 14% |
| Alacchi Moustier | 9/10 | 3987 days | 17% |
| Paul Davies | 8/10 | 3988 days | 17% |
| Maya Bakker-deDreu | 9/10 | 3995 days | 17% |
| Martin Thibaut | 10/10 | 3997 days | 17% |
| Jitka Zenahlikova | 9/10 | 4303 days | 2% |
| Michael Heuer | 10/10 | 4311 days | 2% |
| Eric | 7/10 | 4316 days | 2% |
| Rogier Ramaker | 8/10 | 4316 days | 2% |
| Fraser Goldsmith | 7/10 | 4335 days | 1% |
| Jeanne Moustier | 9/10 | 4352 days | 0% |
| N Zitscher | 10/10 | 4404 days | 23% |
Adjustments
No Adjustment
Several adjustments to the weighted average may be added to improve relevancy and credibility. Haast does not meet the criteria for any of these adjustments to apply.
Balancing Adjustment
2.13% Adjustment
Every experience's review score is adjusted to balance out the disproportional number of negative reviews that are contributed.
You won't be surprised to learn that disgruntled folk are more likely to leave a review than happy ones. They are motivated to share their experience and warn others. We consider this a good thing and it's why reading the reviews is important. However we've learned it can misrepresent the experience in a more overall sense.
We apply a balancing adjustment to counteract this effect and ensure the ranking score is a more fair representation of the experience. This adjustment is applied equally to all experiences.
Final Ranking Score
86%
The final ranking score once any adjustments, ratings, and rounding has been applied. This value is recalculated each day and a short rolling average is applied. Therefore it may not be precisely accurate based on the other values presented.