Simplified integration and usage-based pricing are the buzzwords of Yusp recommendation engine’s eCommerce solution for small and medium businesses. The recent product update includes a new branding and dashboard along with the flexible pricing and smooth integration.
With the launch of Yusp in 2016 Gravity R&D made its enterprise recommendation engine available for smaller eCommerce stores. In 2017 the company has shifted the full scope of enterprise personalization solution under the previously established brand name. Yusp offers solutions in ten business models such as e-commerce, online marketplace, digital media publishing, brick and mortar retail and retail banking.
Since 2018 summer, Yusp’s product portfolio for SME-sized eCommerce companies has been getting shape in a new brand name. ‘Yuspify’ represents the company’s clear focus on leading eCommerce platforms like Shopify. Yuspify opens the gate wide for even the smallest e-commerce stores to try those technology and state-of-the-art algorithms that large e-commerce sites like Cora, eMAG or eBay Turkey have been using for years with outstanding business results. This powerful arsenal is now available for the users of the biggest eCommerce platforms ( WooCommerce, Prestashop and early access for Shopify ) through a simple module integration.
Amazon, eBay, Macy’s or Walmart have been improving their market share with product recommendations for 8-10 years. Yuspify has a key role to democratize the game with its self-developed technology. The available customer service and the importance of its new, performance-based pricing are also crucial steps in this process that ‘levels’ the field between large and small players.
In the following interview, we answer the most important questions about Yuspify with George Herbszt (Head of Product Development at Yuspify) and Mark Milankovich (Digital Marketer at Yuspify) about the product and pricing scheme.
– Where do you place Yuspify in the timeline of recommender system evolution?
New Generation of Recommender Systems is an often used term. It sounds trendy, glossy and sleek, and e-marketers often refer to it with a unique trait: self-service. This term defines the preceding generation of recommender systems that require vasts of human workload, thus small SMEs could not afford it. In addition, there is another immanent feature that new generational recommender systems have adapted: education. The ‘be your own data scientist’ or ‘be your own data miner’ – approach opens an insight window. To sum up, if you are curious to look under the hood, you have the opportunity to immerse yourself in data science.
Our vision is to reduce human workforce need in the second phase of the recommender system lifetime. Namely, a recommender system has 2 principal life phases: the first is the integration, and the second is the aftermath where the self-service characteristic appears and unfolds. To belong to this new generation, we must grab and automate the majority of the personalization demands of our users.
– Did you automate even the plugin-platform integration process? How much effort does it need from the user’s end?
We have already simplified the integration process for eCommerce platforms such as Woocommerce and Prestashop. Here we really provide a plugin that connects the e-commerce store with our system. For Shopify we would like to make things even easier and use the embedded app to create a dashboard available from the Shopify admin interface, where users can also see analytics, change recommendation logics and design.
– The new dashboard of Yuspify simplifies how users can edit and interact with the recommendation algorithms. Can they initiate and implement not only simple but advanced adjustments?
Only the low-risk adjustments. We equip our users with a well-regulated blueprint that they cannot modify but they can set all the important recommendation aspects without having the chance to spoil anything. Our users will be able to set the logics on the complexity level of the following example: ‘displaying 40% higher proportion of red products in the recommendation widgets to those buyers who already bought more than 30% red product.’ We coined these logics to be super safe. Their ‘farming’ and sublimation is also very challenging, not to mention their firm and categoric separation from the fragile logics the users can’t touch.
I would like to shed the light to a very important part of our vision: our client e-stores must have and will have enough extra revenue from the built-in logics that they will consider investing into custom logics that can further boost their KPIs.
– Do you consider closer cooperation with user- or traffic analytics plugin providers whose service can help Yuspify’s performance?
We consider the cooperation with user- or traffic analytics plugin providers because the main point of on-site conversion optimization is to utilize the existing data treasure whose origin could be an external application as well: once you possess the data and the coherency, the rest of the job is only a matter of channel.
– Why are you focusing especially on Shopify with Yuspify?
Shopify is a community technologically homogeneous as well as the three major platforms we are going to focus on (Woocommerce, Prestashop and Magento) Beneath being a recommendation system Yuspify is a data mining tool: the product synchronization and the user tracking grouped together is already a huge result.
– What types of stores can benefit the most from recommender systems?
Entertainment-purpose electronic devices are taking place on the hedge of the product recommendation revolution wave! The algorithms can group these items easily with attributing and sorting their numerous technological parameters. If you have a store in this segment, you can set up tons of interchangeable and cross substitutable product groups. Moreover, because of the diversity, you can capitalize on these offsprings with lots of upselling and cross-selling potentials. These are the grounds where Yuspify can add lots of value.
– At your earlier version, you ordered the pricing into different plans. How much have you moved ahead with the new pricing structure?
We have dropped the old-fashioned pricing that guaranteed the recommendations delivered to the target but not the revenue growth. The new pricing structure bills only after the revenue surplus. Moreover, the former pricing model was not favorable for the smallest users who had to pay 49 USD for the starter recommendation package, but their store maturity was low, and their product catalog was too underdeveloped to imply significant revenue growth even with the smartest recommender engine.
The new pricing of the 30-day trial version is riskless for those potential clients who are just interested in the marketing automatization solutions. The essence of this symbiotic relationship with the users is that we don’t take away more than we give. What does this mean exactly? Even for that user who has a very tight budget, we are still able to serve recommendations.
– How is it possible to measure the revenue growth fostered by the recommendations? How can you separate your recommendation-propelled revenues from other revenues driven by other factors?
With possessing a well-proven KPI measuring, it’s simple: in a given e-commerce store in a given period of time how many purchases have happened through our recommendations.
The essence of these AB-tests:
A. The customer buys through this recommendation in 24 hours
B. The customer hadn’t met the recommendation.
The conversion differences between these two possibilities leave no open questions. See the results of Cora Romania case study where we designed these kinds of A/B tests using millions of visits. The traffic stream was split on a point where half of the visitors was displayed the A page where nobody, half of them was displayed the B page where everybody received recommendations across all the scopes. Users served by Yusp have generated 10% additional revenue during the test period. This knowledge and best practice distilled here is the principal fuel of Yuspify in helping small-size e-commerce stores.
Purchasing by recommendation within 24 hours – this is a KPI we had consolidated before the design process of the SME product started. The point is the completed or non-completed conversion within a certain period of time and the difference of these. Also, we can measure how much the basket value or the visit frequency have grown.
Not only the increased basket value but the more frequently returning user is a guaranteed success. The user is more likely to return if the recommendation engine reduces the basket value with an optimized offer. In an example: we recommend something similar but on a lower price plus the complementary items. What will this visitor remember? This is the store where he can purchase the same product on a lower price.
– Regarding customer support: Wouldn’t it be a problem that smaller e-commerce stores haven’t got enough resources from the support?
We provide all of our users with proportional customer support. Our primary and basic logics (item-, cart-, category-, homepage- logics templates) have enough efficiency to serve smaller and bigger e-commerce stores in an equal way. These simple logics have very low human customer support workforce needs.
No doubt that the smaller e-commerce stores will favorize these rudimentary logics. If a client would like to use sophisticated recommendations in a huge volume, that shop assumably has a bigger size and a beefy budget. A custom recommendation logic always implies surcharges because of the extra support and design cost of the non-repeatable aspects.
– What are the risks of revenue-based pricing?
Losing revenue on our side, if our logics don’t perform properly. That’s why our team targets an immaculate provision of the service that produces the recommendation-generated revenue in 24 hours across 7 days. What kind of risks do e-commerce stores have to face? We should rather put the question this way: what risks will the store say goodbye to? The risk of occurring, unexpected expenses that stores normally can’t forecast. The business can foresee that Yuspify will nab $ 0.03 if $ 1 revenue arrives. The store can pre-plan with the expenditure whose collateral will appear multiplied on the income sheet.