4.3 stars sells more than 4.9 stars
Extremely high average product ratings, between 4.5 and 5 stars, make people skeptical and generate lower sales than ratings between 4 and 4.5 stars.
New to Ariyh? Join 8,719 evidence-based marketers for 3min practical insights 💡 from scientific research 🎓 to get better marketing results 📈
Today’s insight is brought to you by… HubSpot & Mention
The Instagram Engagement Report is a result of an in-depth analysis of 110M+ Instagram posts across 1M+ users globally.
Thanks to key data, trends and expert insights from the teams at Mention and Hubspot, you can optimize your Instagram marketing strategy in 2022 with confidence.
Want to sponsor Ariyh? Here’s all you need to know.
📝 Intro
People trust others’ reviews more than what marketers say.
But they also know that companies can try to hide negative reviews, incentivize positive reviews, or - unethically - plant fake positive reviews. Even more so if the reviews are on the company’s own website (vs on Google or TripAdvisor).
So when a website is splattered exclusively with 5-star reviews, it might be a sign that something is off.
This study from Northwestern University and University of Amsterdam analyzed the effect that different average star ratings had on actual sales of 3 ecommerce retailers.
P.S.: A huge thank you to the 111 of you who pre-ordered The Science-based Playbook of Pricing & Promotions!
It’s a much stronger response than what I hoped for - which makes me even more excited to get it out to you on the 5th of April.
If you haven’t pre-ordered yet, now is a good time. The price will increase +94% 48 hours after launch. You can pre-order your copy here.
Previous insight: $2,111 is better than $1,999 (100+ more insights here)
Review ratings between 4 and 4.5 sell more than those between 4.5 and 5
Impacted metrics: Customer acquisition
Channels: Reviews
For: B2C. Can be tested for B2B
Research date: June 2016
📈 Recommendation
Aim to show an average review rating of between 4 and 4.5 stars, not only perfect 5 stars reviews.
For example, it’s alright to have some negative reviews that lower your average, especially if they are irrelevant (e.g. a bad rating because the item doesn’t deliver to a specific location, not because the product is bad).
Your sales will be higher.
🎓 Findings
Products with an average rating of 4.5 to 5 stars have lower sales than those with between 4 and 4.5 stars.
Average review ratings increase sales as they go up and peak at between 4 and 4.5 stars, then slightly lower them again (it’s still better to have a 5-star rating than a 3 star rating).
This study didn’t find a strong relationship between a high number or reviews and more sales. In fact, for some products (e.g. natural hair care products, herbal vitamins) too many reviews seemed to slightly reduce sales (but the effect was not significant).
🧠 Why it works
For many products, especially if they are low risk (e.g. a small purchase) or when we are mentally overloaded, we don’t dig into the detailed content of reviews. Instead, we rely on what the average rating is.
If the average rating is extremely high, we tend to think that they are too good to be true. We suspect that the company could have manipulated them.
Instead, when some negative reviews are present but the rating is still good (4 to 4.5 stars), we assume that the reviews honestly discussed both positive and negative aspects, but the product still came out as great.
✋ Limitations
It’s unclear whether it’s better to have 10 good (e.g. 4-star) reviews or mostly excellent reviews (e.g. 5-star) and a handful of negative ones. Based on previous research, it might be the latter.
This study focused exclusively on reviews shown on companies’ own websites, where reviews might be more easily manipulated. People might have a different reaction to extremely high ratings on 3rd party platforms (e.g. Trustpilot, TripAdvisor), where it’s harder to manipulate them.
The retailers in the study sold different types of products (e.g. light bulbs, women's athletic shoes), but these were all physical products. It’s unclear whether the effect stands for online products, services, or more expensive items.
🏢 Companies using this
Typically, companies follow one or a mix of the following approaches:
Aim to show the highest possible rating, sometimes incentivizing positive reviews
Try to suppress or hide negative reviews
Hands-off, paying little attention to review ratings
When companies highlight star ratings of reviews on their homepages (e.g. Basecamp) they tend to only show 5-star ratings.
⚡ Steps to implement
When you have great reviews, don’t try to hide or censor a few negative reviews that will average things out, unless it’s your first review and it’s negative.
If you constantly receive low ratings, improve your product or service, or stop selling it.
Show good reviews on your homepage, especially if they address customers’ main objections, and experiment with showing them in ads or in other channels (e.g. email).
🔍 Study type
Market observation (analysis of sales of 3 ecommerce retailers between 29 Jun 2014 and 11 October 2014)
📖 Research
Too good to be true: the role of online reviews’ features in probability to buy. International Journal of Advertising (June 2016).
🏫 Researchers
Ewa Masłowska. University of Illinois Urbana-Champaign
Edward Malthouse. Northwestern University
Stefan Bernritter. King's Business School, King's College London
Remember: This is a scientific discovery. In the future it will probably be better understood and could even be proven wrong (that’s how science works). It may also not be generalizable to your situation. If it’s a risky change, always test it on a small scale before rolling it out widely.
Rate today’s insight to help me make Ariyh's next insights 🎓 even more useful 📈
How was today’s insight?
Loved it | Great | Good | Meh | Bad
What else you can get out of Ariyh:
📣 Want to advertise on Ariyh? Here’s all you need to know
📈 Want a quick boost to your business? Ask for 1:1 personalized recommendations here
New to Ariyh? -> Subscribe below or read other 3min marketing insights here
This is based on just 3 companies, so I would be wary of any findings. Even if the total # of observations is large, there might be bias from the company geo/vertical/demo etc.