Rethinking Feedback for Product Teams - Moving Beyond Tagging and Transcripts

The path from feedback to insights

I've been working in product teams for the last 7 years, and one thing I've repeatedly noticed is the sheer scale of feedback that's around. Whether that's from user interviews, user tickets, chats, NPS scores, or surveys. More often than not, I've found that the issue isn't getting hold of the feedback, it's doing something useful with it. Most if it, in my experience, ends up being archived, skim-read, or tagged for later - but actually acting on the feedback can be really tricky.

Some systems, like Dovetail, help with this, in particular with storing and "coding", or tagging the data, which can help stay on top of the data that's inbound - but doesn't go far enough in actually having an impact on teams.

How we got here

These sorts of platforms work well for really focussed, "deep-dive" UX research on a hyper-specific functionality or feature, but fall short when it comes to holistic understanding and action of CX data.

Many of these platforms are exploring new AI technologies, like "Magic Summarize", which definitely can help in certain situations, but due to their naïve approach, can lead to hallucinations or more manual checking, which only adds to the workload.

This all means that working directly with a user interview videos, and quite strict taxonomies works well with these tools, but sadly, that often doesn't reflect the full state of user feedback. Lots of necessary detail gets left by the side - data which could all help in building better, more useful products, and delivering value to users.

Shifting forwards

In today's new, hyper-focussed industry, what matters most is clarity, impact and speed. Producing research decks can be useful, but ultimately we all want to ensure that we're improving the products we ship to users, through properly listening and acting on what they're asking.

This inherently means moving from the big scale (quarterly, yearly interviews) to the small, continuous scale - the companies who are innovating the most are responding to weekly product updates, seeing how user tickets correlate directly to churn, and tracking the impact of live performance campaigns.

For feedback tools to work effectively, they need to surface what matters, who should act on it, and what the impact on the wider business will be. And this all needs to happen at scale, with as diverse a range of data sources as possible. This gives the best possible picture of what's going on.

A picture of next-generation feedback

From what was discussed above, it's then clear that in order to have the biggest impact, any tool needs to:

• Work at scale: it should be input data agnostic, work efficiently with a range of formats and inputs. You don't want to spend hours wrangling your data into the correct format to get results.
• Be well structured: ensure that themes and topics are flexible, and reflect how teams and products operate - without hours of manual coding, and work more deeply than keyword matching
• Be speedy: using "always-on" data mechanisms, rather than discrete, manually triggered studies.
• Tell stories: pull together insights that delivers on what's being prioritised at a company level, and not just an average synthesis of the data sources.

We're trying to build that tool at Sunbeam.

Real problems

Some examples we've seen product teams trying to solve are:

• "Why are bookings down?" - attempting to tie in specific causes, such as app friction, or support wait-times to concrete outcomes.
• "What factors are driving churn?" - highlighting a key cluster of complaints right of the back of pricing changes.

These are real problems that users have, which the data can point to, if it can be ingested, processed and delivered to the right person at the right time.

Why now?

Especially now, more than ever, expectations for product teams are higher. There's faster product cycles, and customers expect more from the companies they engage with. Whether that's strong, consistent customer support, to direct product improvements, it's all important.

As mentioned above, tools like Dovetail work for research snapshots, but really don't function well at scale - the technical limits (1,000 insights, notes or highlights) just don't work for the scale that we're talking about. These don't work to keep up with user experiences as they're unfolding.

Teams directly need clear, timely insights that speak to business outcomes. Outcomes that can shift and change.

Building a new culture around feedback for product teams

We strongly believe that collecting more feedback isn't the answer - we all have enough feedback. The answer lies around making what we have more useful. That means moving beyond only involving researchers to making whole product-teams "feedback first". We want to empower everyone, from operations, to product, to support, to have access to actionable insights, that align with company goals.

We really believe the future of feedback needs to be dynamic, built for scale and for operational impact. This not only comes from tooling, but shifting the whole culture around feedback in product teams.

At Sunbeam, we're trying to build the future of feedback. If any of the above resonated with you, let us know. If you're encountering some of the problems discussed above, we'd love to chat. You can email us at hello@sunbeam.cx, or get in touch via our website.