A recent meta-analysis of 6700 experiments on large ecommerce sites was carried out by Qubit Digital LTD. This study was from the retail and travel sectors, grouping together common treatment types performed on websites. Qubit found that cosmetic changes, in general, have a much smaller impact on revenue than techniques grounded in behavioural psychology.
The 5 methods that were the most successful:
- Social Proof
- Abandonment recovery
- Product recommendations
Let’s take a look at each of these methods, in turn, to see how and why experimenting with them could boost conversion on most ecommerce sites.
The data compiled by Qubit showed that scarcity had an uplift probability of 84% overall.
You will most likely have come across this technique on eCommerce sites such as John Lewis, Amazon and Argos. The quantity left in stock is displayed and in some instances highlighted by a red, amber, green colour scale to indicate numbers available.
Given that ‘scarcity’ labelling plays on our innate FOMO (fear of missing out), it’s not surprising that this method results in increased sales conversion when used. Once someone is browsing your site it’s likely they have a desire to purchase something and so, FOMO techniques like this help encourage them to click that ‘buy it now’ button. Scarcity increases the desire for that particular item, summed up perfectly by changing minds.
“If something is not scarce, then it is not desired or valued that much. Praises from a teacher who seldom praises are valued more than praises from a teacher who is liberal with his or her praise.
Scarcity is a non-linear process. As something becomes more scarce or less scarce, the desire for it does not change in a proportionate way.
If everything is scarce, then scarcity itself lacks its value and people become too used to it. Studies of retail sales have shown that if more than about 30% of goods have ‘sale’ sticker on them, the effectiveness of this method decreases.”
I strongly agree with the last point in the above quote. Scarcity doesn’t always work if it is not used in the correct way. If the scarcity is clearly BS and your customers are smart and see through it, it’s going to do more harm than good.
Social proof is another technique which is increasingly popular on ecommerce sites and well worth experimenting with. Argos and Amazon use it on their product pages, for example. The data compiled by Qubit also showed that social proof had an uplift probability of 82%.
Social proof is, in a way “crowdsourcing” your customers to achieve higher conversion through product reviews, feedback and Q&A’s. Product reviews always have an impact on the potential customer viewing the item for sale because simply put some customers will trust other customers more than they do official descriptions.
For example, if a product has 500+ 4-star reviews, this builds trust in the product and in turn your company because people often comment about the seller in a product review. Amazon also has a product Q&A feature with search functionality (customers helping other customers, a great example of social proof and as a side effect it saves Amazon money in customer service queries).
Quick! Buy this now! Qubit showed that urgency had an uplift probability of 70%.
Put simply urgency messaging plays on “peer pressure” and the desire to get an item before it runs out. What happens? You load a product page and immediately see pop-ups designed to evoke thoughts like “what if it goes out of stock?” “what if I don’t get it in time?”.
Just like the scarcity technique, this plays to our ‘fear of missing out’ and when urgency is combined with a low stock message, it can be exceptionally effective.
Often, companies will also use delivery timescales and countdown tools to achieve similar results. Displaying “delivery available tomorrow if you order within the next 2 hours” or “free delivery for the first 100 buyers” can often be the push needed to convert immediately.
Abandonment is an issue for all eCommerce websites. At some point, your customer will have something that distracts them from purchasing or they might simply change their mind. While this can’t always be prevented, there are methods that can be adopted to help counter it. The data compiled by Qubit showed that abandonment recovery had an uplift probability of 71%.
This actually has a simple solution, have free delivery or free if you spend over £X amount. If you really need to charge for postage then be upfront about it. Have it prominently displayed on your website before a customer even starts to browse. At Argos for example, you can immediately see that you can pay £3.95 for same day delivery or free collection from a store.
The second reason (and a personal bugbear of mine) is when users are required to sign up for an account simply to complete a purchase. I can’t put this simply enough: STOP. DOING. THIS. To state the obvious, don't force a first-time buyer to create an account before they can complete their order. Offer guest checkout instead.
By simply moving the (now optional) registration process to after purchase, users are actually more likely to sign up. Plus, registration is no longer a pinch point for low conversion rates.
A well-placed product recommendation can be the difference between a customer abandoning or converting on most ecommerce sites. The data compiled by Qubit showed that product recommendations had an uplift probability of 76%.
Have you ever made a selection in a local shop, based on a recommendation from the owner or shop assistant? Have your friends or family made recommendations which have been the deciding factor in your purchase choice? Some customers only have an idea of what they are looking for - “I want to buy a 3 seater sofa” but aren’t sure on what type, design, material etc.
The customer will search for a 3 seater sofa and click the first one which looks more or less what they are thinking about. This is the point where the collaborative-filtering method comes in. It incorporates data from users who have made similar choices, then combines that information to make decisions about recommendations.
You can also use content-based filtering which collects data about the likes and dislikes of each visitor (using cookies for tracking over multiple visits), then makes recommendations based on past choices by that user.