Role: Product Design Lead
Company: Ometria
Year: 2020 (Project still ongoing)
Problem statement
Clients do not trust current solution due to the lack of control they have over what it’s shown to contacts. Lack of clarity about the product recommendations logic leads to inconclusive results and some users decided not to use product recommendations or never experiment with them. E.g A/B test. This leads to revenue loss as product recommendations could lead to a higher conversion rate and uplift in revenue.
Background
Clients have been very vocal about the lack of functionality for product recommendations e.g OR logic instead of AND, inactive and out of stock products are shown, the fallback model is very limited and leads to show very generic products in very personalised/targeted messages.
Objectives
Improve the existing experience, make recommendation building intuitive that needs minimal hand-holding and troubleshooting
Provide industry-standard product recommendations that will lead to customers satisfaction and help the sales process.
Definition of success
100% of clients who stopped using the feature for some use cases will adopt product recommendations in new campaigns
Clients interviews
I worked closely with the product owner and customer success manager that facilitate the interview.
7 client interviews have been performed to 4 ideal customer profile and 3 Acceptable customer profile that stopped using the product recommendation
Customers joined with multiple stakeholders, mostly CRMS and Customer Experience managers
Discovery
Missing AND/OR logic preventing them to create efficient recommendations
Excluding out of stock and inactive products is one of the most important features for retailers
Lack of reporting prevented retailers using the engine, as they didn’t have the means to attribute the revenue to the recommendations engine
Not able to see how the recommendations would like was a show stopper for some clients as they really cared about their brand integrity and the marketing they send to the customers were carefully designed. They really wanted to see the recommendations visually to have total control.
Some of the users interpreted wrong the current UI, there was some confusion around the functionalities
User journey & concept
I was given enough information from the existing journey and the discovery to visualize the new journey's structure. Early decisions were:
Add the missing features to the flow that clients flagged in the interview process.
Group and breakdown components in 2 categories: Initial setup (same for each recommendation engine) and engine setup (changes depending on the engine selected in the previous step
Add a preview of the results of the recommendation. Users would be able to see the impact of their choices in real-time.
Use the new Ometria design systems (React) for faster integration.
Interaction design
Outcomes
We explored 3 ideas with the stakeholders and around 12 versions of prototypes
User testing of the final version truly showed the user delight. Clients were happy with the additional features and completed the given tasks with ease.
Recommendation engine reporting was the most successful new feature. Lack of visibility was creating trust issues and was the most significant friction. Recommendation reporting allowed customers to experiment more with the feature. Therefore they could generate better recommendations and drive more revenues.
4 customers currently beta testing the feature. It will be released end of January for the rest of the customers.