42x shorter time-to-insight with AI-first research methods

Butternut Box is the UK's leading fresh dog food company. They deliver freshly cooked dog food, made with human-grade ingredients, to thousands of happy customers across the UK.

London, UK


Build a continuous product discovery practice to quickly validate opportunities and prioritise their product roadmap.


Identifying product opportunities and getting continuous product feedback was manual and didn't scale.


Shortened the time it takes to get user insights from two weeks to a few hours with in-product AI-led user interviews and concept testing.

Butternut Box is a rapidly growing UK-based pet food company that delivers fresh, healthy, and personalized meals for dogs. Their subscription-based service caters to dog owners seeking convenient and nutritious pet food options tailored to their pets' specific needs.

The well-being of dogs and their owners lies at the heart of what the team at Butternut Box does, so involving customers in their product research plays a crucial role in delivering a great customer experience and fueling growth.

To continue creating products that meet the needs of their users (and their dogs), the product teams at Butternut Box strive to understand what their customers experience when using their product.

Challenges obtaining continuous user insights

Recognizing the importance of hearing directly from their users to help prioritize customer needs and improve their products, Butternut Box aimed to do user research throughout the product development lifecycle. But understanding what customers wanted wasn’t easy:

It took weeks to get customer feedback

Setting up user interviews or collecting product feedback required Butternut Box to use their CRM tools to email their customer base in bulk. This often demanded coordination with multiple departments such as analytics and the CRM and marketing team to procure customer lists and send out email blasts. Due to competing projects and deadlines, this process typically took 14 days and required significant back-and-forth to identify the right customer lists, set up the email sends in the CRM tools and liaising with their customers once they responded.

Feedback they collected wasn’t relevant

Identifying customers to invite to their research studies required significant coordination to select users from specific segments. The necessary user segmentation was often unattainable in their CRM tool, hindering Butternut Box from reaching out to the necessary customer segments.

Additionally, the feedback received through their email blasts often wasn't relevant to the features and tasks they were researching, and they often had to question participants about actions taken weeks prior, or arrange mock scenarios where participants were quizzed about hypothetical test situations.

Validating new features and ideas was cumbersome

The challenge of obtaining relevant feedback from customers throughout their development process made it difficult to do customer discovery continuously. The effort necessary to gain user insights about how to optimize their product journeys and which features to prioritize meant that new features often weren't tested with real users until after they were launched. As a result, these features frequently failed to impact their acquisition and retention KPIs, stunting growth.

Building a Continuous Discovery Habit with In-Product User Insights

Recognising the limitations of their current setup, Butternut Box sought more effective methods to quickly and continuously glean user insights directly from their customers. Their primary objective was to establish a habit of continuous, rapid opportunity discovery, where customer feedback informed every product decision. Here's how Butternut Box began constructing their continuous discovery habit:

Integrated Wondering into their product to target customer segments in real-time

To access customer segments based on user data and behaviors in their product in real-time, Butternut Box integrated Wondering's in-product user insights platform into their web application. This allowed them to specify the user behaviors they wished to investigate and deploy AI-led user interviews in-product, and ask users to complete recorded concept tests directly with those users. They could now reach users previously difficult to identify, obtaining user insights and feedback in real-time.

Deployed in-product AI-led user interviews to swiftly pinpoint product opportunities

The product teams at Butternut Box no longer needed to plan weeks in advance each time they wanted to incorporate customer feedback into their design and prioritization process. This enabled them to send out in-product AI-led user interviews with customers across their core product journeys.

On average, within 8 hours of launching their studies, the product teams at Butternut Box received targeted feedback from, and conducted exploratory interviews with, real customers at core product journeys.

To ensure research studies were only sent to the right customers and didn't overwhelm their users, Butternut Box used Wondering's built-in rate-limiting and in-product targeting. These ensured studies only reached a subset of users, and the same user didn't encounter multiple studies within a short time frame.

Used their study findings to continuously prioritize their roadmap and identify new product experiments

The real-time feedback obtained from their in-product surveys and user interviews enabled the product team to make informed product decisions more quickly. Now, instead of waiting until after the launch of a new feature to evaluate its impact on customers, they could refine and prioritize upcoming launches confidently, knowing their decisions were backed by insights from real users.

The impact: Reduced their time-to-user insight by 42x

Like many modern product teams, the major impediment preventing Butternut Box from establishing a continuous discovery motion was the effort required to coordinate continuous feedback collection from customers.

Having continuous access to feedback and insights from their users through Wondering's user insights platform, the product teams at Butternut Box can now deploy in-product concept tests, conduct real-time user interviews, and measure product satisfaction through with customers across their core product journeys.

Before incorporating in-product research studies into their continuous discovery process, coordinating the collection of insights through email surveys and organizing user interviews would take an average of 14 days. By utilizing in-product user insights through Wondering, this now takes the team an average of just 8 hours, representing a 42x reduction in the time it takes to obtain user insights.

This proved especially valuable during the challenges presented by the global pandemic and the ensuing economic instability. Butternut Box was able to continue adapting to their market and customer needs, even as their users' needs and preferences changed. Now, re-testing their assumptions and obtaining fresh data through in-product research studies enable them to glean new insights continuously within just a few hours.

Accelerating your time-to-insight

In-context user research studies such as in-product AI-led user interviews and concept tests are excellent ways to collect targeted feedback and user insights from your actual customers. They can help you:

  1. Optimize your product journeys: Uncover pain points and areas for improvement in your core product journeys.
  2. Validate new ideas and test your designs: Test your designs before building, and swiftly validate new ideas and concepts.
  3. Prioritize your product roadmap: Make informed decisions about which features to build next, based on a deeper understanding of your users.

Wondering empowered researchers at both large and small companies to increase the success of their products by delivering relevant user insights in real-time from real customers. Try out Wondering for free to start scaling your user insights programme.