How analytics curves more easy to understand?
State of playKolsquare is a Saas B2B platform dedicated to help professionals optimize their Marketing Influence campaigns on 5 social networks (Instagram, Youtube, Facebook, Twitter, TikTok).
An important part of this platform is the influencer description. On this section, a user can find numerous informations about the influencer' profile, audience, statistics, content, interactions with brands and social networks users...
The analytic section is one the oldest created and needed a real update. It's composed of two part: a table displaying basic calculations per networks (average engagment rate per post, number of post per week, views likes comment or shares per posts...). The second part is about visualisation of the performance with curves. These curves are the topic of this case study.
Above: performance curves.
Main Issue(s)No discovery on these performance curves has been made, because you will see there are many non-sense with these curves, and it's all about optimizations in the run process of our tech sprints.
- Issue #1: Scales
- Issue #2: Missing datas
As the curves are presenting datas, some of them can be missing. There's a lot a reasons that lead us to have incomplete curves. For example, when a network is changing its rules for detecting bots, when a user switch from a public profile to a private one, or simply because we're not able to rewrite the past, and when we begin collecting datas from a profile, we can't go back and retrace the whole evolution of its community, or engagment rate.
- Issue #3: Post & stories
Stories are an important part in the influencer game, and when the influencers began to massively use stories, our product was not prepared. The informations we gave were only posts informations. Stories needed to be added.
- Issue #4: Lack of added value in the given datas
When hovering the curves, a tooltip shows informations of the day you are hovering. But the way these informations are displayed makes them totally unreadable, and so useless.
- Issue #5: No explanations
Tooltips are unclear. When there are tooltips. :)
- Issue #6: Unprecise date selection
Currently, it's not possible to select precisely an amplitude of time from date to date, which could be really useful when checking your competitor's campaigns or following the activity of a specific influencer.
Above: issues illustrated.
Identified solutionsAs this project is an unasked improvement, it hasn't been submitted to users yet. But it is has been to the product and tech teams, that validated its feasability.
Above: performance curves module considered.
To solve the overall issues, here's what I've found:
- Issue #1: Scales and unprecises date selection
To avoid viewing gaps, the community curve is divided in two to be put in perspective: a long-term vision, and a short-term vision with selectable dates.
- Issue #2: Missing datas (view the screen hereafter)
When datas are missing, we're just not displaying anothing, and clearly announce the reason why these datas can't be seen. For exemple, when the account has been switched to private, or not collected by our teams.
- Issue #3: Posts & Stories
Posts and stories are now displayed in the publication section.
- Issue #4: Lack of added value
Besides the tooltips, from now on, a sidebar is giving precise informations depending of the dates selected. On this sidebar, the KPIs can be different depending on social feeds, and on the main content area, information displayed can be added. For exemple, for youtube especially, community does not make any sense, but number of views does.
- Issue #6: Unprecise date selection
From now, you can select via the datepicker, or directly on the first curve a precise date range, that will have an impact on the analytics on the sidebar, and on the following curves. In case of lack of data, for exemple if we're collecting an influencer' feed since only 5 days, I think this time range must be disabled under a fortnight.
Above: an account with only few days of history.
Above: analytics on another social feed differs.