This is the story of how we developed a feed leveraging time-series algorithms to help brand managers and PR teams discover emerging conversations.

My role:

Lead product designer

Designing for:

Brand managers / Industry researchers / PR teams

Deliverables:

Wireframes for testing / User-interface designs / High-fidelity prototype / Handovers for developers

The result

A live events feed

A feed that explored trending conversations and helped users understand their main drivers

Pivoting to meet the users needs

Pivoting on the back of research to ensure we delivered  what users needed

Leveraging existing technology 

Using a time-series algorithm to surface changes in the data over time

The problem

Brandwatch Consumer Research is complex and time consuming. Novice users do not have the time or expertise to set up and explore in-depth dashboards to find insights. 

In order to forgo some of the technical constraints that Brandwatch Consumer Research (BCR) currently presents, the product team started to explore the possibility of a stand alone go-to-market product. The assumption was, that expert users could turn to BCR to dig deeper into insights while the light version would allow novice users to get top level insights on their brands and industries.

Unpacking the stand alone product 

We started by unpacking the use cases and user stories that we had received from the customer experience team. This helped us develop themes that the user desired  and were in line with the product vision. 


These user stories and conversations helped us inform the information hierarchy, which in turn helped us establish a product. The stand alone product allowed users to set up quickly, gain insights about their brand and industry, view any emerging conversations, and track changes over time. 

By collaborating with the engineers, we worked on the events feed which was the core part of the product. The feed was created from a time-series algorithm to analyse distinct drivers over time to identify events. The threshold of the events are high, medium and low depending on what the data suggested showed over time. If there is a shift in the data, then it appears on the user's feed. 

We also used AI to generate concise and informative chart titles to summarise what the data presented  in the charts.

How did users resonate with the stand alone product 

We then set out to test with clients and internal users (Customer Success Managers and Account Managers). Here’s what the research uncovered. 

Pivoting to fit users' needs

Our research concluded that having a basic stand-alone product did not give users enough context regarding their brand or industry. Users wanted the simplicity of setting up their data, but with the granularity that BCR offers. Users found the events feed and the AI-generated chart titles were the most appealing features of the product. 

We decided to pivot and not launch a product that doesn't meet user expectations and needs. Instead, it was decided to focus on how we can bring the events feed into the existing product and share our insights on the AI-generated chart titles with another team who can work on it.

Navigating tech and design challenges

Focusing on launching the events feed in the legacy product came with its own challenges. Here’s how we tackled them. 


1. Events can be categorised with a ‘high, medium or low’ threshold

The engineers very quickly discovered that the low priority events were mostly spam, therefore we decided to not show events categorised as ‘low’. To distinguish events with a high threshold we decided to mark them with a most relevant tag. Also, by having them appear first, users know which ones are the most trending.

2. The chronology of events 

By showing high priority events from newest to oldest followed by the medium priority events, users were able to see the most pressing conversations as they happened. 

3. Visually distinguishing the events

The engineering team sought to discover how to pull images for trending articles, tweets, authors and hashtags. However, we were unable to surface these. Therefore we settled on using icons and visual cues to help distinguish the events from each other. 

The learnings

Pivoting to user needs

Communicating that a stakeholder's vision may not be the most desirable path was challenging. However, prioritising user needs was crucial, as we felt it was essential to pivot our direction based on the valuable insights we gained from user research.


Consult with engineers as soon as possible 

While creativity is key, a big learning for us was working closely with engineers and understanding what can and can’t be done early on.