Strategies For Better User Reviews
In ecommerce, a tried and tested method of improving conversion is to provide information on "what other people think about this product". (User reviews).
As well as influencing conversion at the point of sale, they can encourage customers to click on your site in the first place. How? By showing ratings in your Adwords campaigns.
Back in 2007, I wrote a blog post entitled Web 2.1 - Reviews From People Like Me.
I wrote,
User reviews need to evolve to enable customers to filter reviews so that they can see reviews just from "people like me".
Reviews like this did not exist in 2007. They were generic, a one-size fits all approach. I'm glad to see that since then we've moved on.
The best example I've seen to date is Booking.com (for hotels). In travel, group size and type is probably the most defining attribute, so this is a useful way to make the reviews more relevant.
Bookatable.com (for restaurants) do it differently, using tags. It's a different approach, but it gives a sense of the type of occasion that that restaurant would be suitable for. For dining, occasion is the most defining attribute.
Kia cars have a reviews section. Here the defining attribute is the driver profile; commuter, motability, motorway driver, parent, recreational or other. Note how Kia reviews are from drivers who have a confirmed purchased; these are credible reviewers. (Compare this with TripAdvisor where anyone can review a hotel and you may never know if they actually stayed there or not).
Amazon however still have a generic system. A reader of a book perhaps does not have a defining attribute and perhaps books themselves don't lend themselves to more in depth review categorisation.
Segmentation matters, some of the time
A parent will judge a hotel on different criteria to a group of friends.
A business lunch may work well in a restaurant yet the same restaurant might not be suitable for a romantic occasion.
A commuter values a car differently to a high mileage motorway driver.
Yet a book? That's less dependant on a category of buyer.
One person's apple is another person's poison, so goes the saying.
Here's the key point; Segmented reviews are most useful if the buyer's profile is likely to affect the enjoyment of the product or a service.
In 2007 I wrote
This is particularly important in the case of experiences. Experiences include things like destinations, hotels, restaurants, theatre, travel and other services. The use of a tangible product (e.g. hairdryer, camera) is less affected by personal preferences. With experiences, preferences do count. Peer reviews can actually be misleading rather than assisting.
From the same year, Marketing Pilgrim reported, "Almost 50% of online shoppers find user reviews helpful. That 50% spends more online than the other 50%."
Clearly reviews are important in ecommerce. Smart retailers are segmenting their reviews to become more useful to buyers.
A note for startups
Being able to segment reviews by user behaviour type requires sufficient volumes of transactions. Whilst you might not be in a position to provide segmented reviews from day one, it's never too early to capture the data.
How many reviews do you need to make segmentation meaningful? Common sense would suggest that you need to have a double-digit count of reviews (at a product level) for a filter to make sense.
You can always build the input right from day one (= a collection process that gathers the right data so you can segment at a later date). The output can come later (displaying segmented reviews).
Takeaways;
- Reviews are important for many (but not all) buyers
- Segmented reviews work best where the product/service experience can be affected by the profile of the buyer
- To segment well requires gathering the right data to start with. A good output requires a good input
- It's worth considering a volume threshold to only segment reviews at a product level once you have sufficient reviews for that product to make it meaningful
- A good model is to allow users to filter on a primary category and then show a variety of scores for attributes of the product that vary according to the filter used (e.g. Booking.com)
- Reviews from confirmed purchasers are more authentic and trustworthy