FAQs – CSAT

Let us focus on the 3 most commonly used scales for customer satisfaction:

• CSAT measures satisfaction with the product/service on a 1-5 scale. Reporting the metric can be done through the mean of all ratings across the customer base, or looking at the aggregate % of respondents who provide the top 2 rating, which provides a wider range compared to the mean score. CSAT is quantifiable and easily understood, however it is more focused on short term satisfaction.
• NPS is the most widely used metric, due to it being a proven indicator of customer satisfaction and loyalty. NPS asks customers how likely they will be to recommend the company to family and friends on a 0-10 scale. Based on the responses, customers can be bucketed into promoters (Rating 9-10), passives (7-8) and detractors (0-6). NPS is then computed as the difference between % promoters and % of detractors. NPS can be used in the context of the transaction (Transactional NPS) or the overall relationship with the company (Relationship NPS). NPS is widely popular and used by many companies, and it is interpretable by every stakeholder.
• CES is built on the premise that ease of the experience is the most important driver of customer experience. It is typically measured on a 1-5 or 1-7 scale and asks customers to rate how easy the company made it for them to complete the action. While CES may be useful in some contexts than the others (Call centre interaction experience, for example), the premise that ease of the experience drives customer satisfaction is not common across all interactions.

While each scale has its own benefits and costs attached, it’s really not about the scale, but what you do with it that matters. A scale that can classify customers into happy and unhappy ones is typically enough. What’s more important is the exercise of satisfying your unhappy customers while monetizing your happy ones.

NPS by design is a more spread-out scale. It makes no distinction between a detractor who rates you a 6 and a detractor who rates you a 0.

The intent of NPS is to spread out the customers who would have probably given you a 3 or 4 on the CSAT scale and identify the detractors and passives properly. It is not uncommon for businesses to get a low NPS score on the first attempt. A low score in itself isn’t a problem, but a low score that persists overtime is. The NPS only serves as a north-star to the continuous improvement journey.

Businesses must focus less on the score and more on the improvement of the score in hand to begin the journey.

Comparing your transactional and relationship NPS/CSAT scores with those of your competitors is not a like-to-like comparison. This is because your scores are based on the ratings of your customers, while your competitor’s scores are based on the ratings of their customers.

Also, there might be marked differences in the survey methodology adopted by different companies. The only way benchmarking can work is that the study is conducted by a neutral third-party on a common customer base representative of the population with a standard methodology.

For a 1-5 scale, tracking the variations in mean score over time has the following drawbacks:

• The variations are not dramatic (e.g a 10% increase from a CSAT of 3 is 3.3, a 0.3-point increase).
• The mean does not take into account the distribution of ratings (e.g You can get an average of 3 when everyone gives your brand a score of 3. You can also get an average of 3 when half the respondents give you 5, while the other half gives you 1. Both scenarios are remarkably different in context).

In contrast, reporting the Top 2% or Bottom 2% gives you a wider scale for variation (0-100%). Viewing these metrics together also lets you put your rating in context.

0 to 10 is only a scale. It is what you do with the scale that makes it NPS. So you could hypothetically use this scale and calculate ‘mean’ as the score, or consider the top 2% ratings/bottom 2% ratings. Conversely, even a 1-5 scale can be converted to NPS by subtracting the % of respondents who rated 1, 2 or 3 from the % of respondents who rated 5.