PROJECT

Collaborative conversation design
Improving how customers collaborate on chatbot conversation design
Company
Drift
Company
April 2022
Overview
Problem
The learning curve of the chatbot builder and paid seat requirements were prohibitive to collaboration conversation design.
Solution
Give collaborators access and collaboration capabilities for conversation design.
Results
Product-market-fit achieved*, based on 50% of customers reporting that they'd be "very disappointed" if they no longer had access to the feature.
* A score of 40% or higher is considered to have reached product market fit
DETAILS
Understanding the need for better collaboration
In 2022, challenges associated with collaborating with others on chatbot content was one of the highest ranked customer issues. And that's because gathering feedback on chatbot content with colleagues, a critical step in the conversation design process, was incredibly time-consuming, if not nearly impossible, to complete. Over and over we heard customers say things like, “just build a way to export a pdf of the builder, that’ll solve my problem,” but our team knew it wasn’t as simple as that. I decided to do some customer research to unpack this problem a bit more. My goal was to understand how customers collaborate with others on chatbots today, identify the biggest pain points with the current process, and gather high-level feedback on some initial concepts.
Customer learnings
Through an initial survey, and then a series of customer interviews, I learned that many customers were reviewing chatbot content with their colleagues synchronously (e.g. requiring a meeting). And not because they wanted to, but because they had to. Even if their colleagues had access to the chatbot builder (which the majority did not due to the additional cost of seats) the interface was too foreign and too overwhelming for folks to conceptualize what they were looking at. It was designed for a power user, focusing on additional logic and behind-the-scenes marketing automation that most reviewers didn't need to think about.
To avoid paying for additional licenses/seats in Drift, and to visualize their conversation flows in a more user-friendly way, a common workaround was to rebuild their chatbot flows in another tool (Miro, Figma, Google Diagrams). Using these tools had the additional benefit that they were commenting/collaboration tools that folks were already familiar with. But these workarounds brought about other challenges. It was double the amount of work for conversation designers to have to build their chat flows in another tool, collect feedback, and then rebuild them in Drift. And unless they were able to continuously update both what was in Drift and their external collaboration tool, the chatbot versions were never guaranteed to be in sync with one another. It was a major headache for customers.
In addition to interviewing customers about their current challenges, I tested a number of concepts on the customer calls. Randomizing the order for each interview, the concepts included ideas that customers had requested (visual chatbot export) to a full redesign of the chatbot interface specifically for reviewers. These concepts proved extremely useful in facilitating discussion about the pros/cons of different approaches, as well as a means of uncovering more details of their collaboration needs. For example, our assumption that being able to view an always up-to-date chatbot version was far superior to sending around a pdf that’d become outdated quickly was validated through these discussions. I also learned that there was value in improving the interface to make it easier to read, even without commenting capabilities (though this was definitely our eventual target.)
A focus on time to value
To solve all of these problems at once would be a large design and technical lift, which would’ve meant that it’d take quite some time before it was in customer’s hands. Wanting to provide value to our customers as soon as possible, I worked alongside my PM and Engineering teammates to put together a plan for tackling these issues incrementally, starting with the general viewing improvements.
Measuring success
We launched the first incremental improvement to a beta group early summer 2023 and, pending the completion of features that were scoped out due to shifting team priorities, will be released to all Drift customers in December 2023. Usage amongst the selected companies and their non-license-holding teammates has steadily increased since the feature's release, and the feature has a 45% product-market-fit rating collected via in-product survey.