One of the three versions of the dashboard we prototyped to test.

Goal

When gathering NPS data for our experience we noticed a trend in the verbatim responses that quite a few users who responded to our NPS response had no idea that they even had an instance of our application. The hypothesis I formed was focused on two possibilities:

  • People couldn’t find our product even though it was included in a cloud foundry application on IBM’s public cloud.

    • or

  • People could find it but didn’t understand what it was.

I then built out a multi-variate test to check those hypotheses

Method

Run a Split test of different versions of the IBM Cloud unified dashboard that their team was working on to see how they effected our product.

  • 3 Versions of the dashboard

  • 30 Participants

    • A > B (5 Participants)

    • A > C (5 Participants)

    • B > A (5 Participants)

    • B > C (5 Participants)

    • C > A (5 Participants)

    • C > B (5 Participants)

  • Data To gather

    • Task Completion Rate (Self Reported)

    • Task Completion Rate (Observed)

    • Ease of Use

    • Usefulness

    • NPS

    • Preferred Version

    • Qualitative Data

 

Sample of the 10 Excel pages of data we collected

Impact

After presenting the deck below at a VP level meeting I was able to get buy in on reassessing some of the Cloud teams decisions. The insights created from this study formed the basis of a months long collaboration between the our team and theirs.

Lessons Learned

I think the key to the success of the study was testing designs that the Cloud team was already considering. If we had built our own dashboards I think it might have fallen on deaf ears but by collaborating with the team already working on it we were able to tell a consistent story.

Presentation Deck