This case outlines how Astutus Analytics helped a start up community group clarify it's members expectations in a limited time window using screen scraping technology and a little bit of database know how.
Names have been changed but the process and the numbers are all accurate.
Dave had two passions in life: He loves going to museums and he loves to tinker with electronics. Recently Dave discovered meetup.com, a website that allows you to promote special interest groups in your community. Immediately Dave saw this platform as a great way to meet other people with his passion and so he started a new group, People Hacking Libraries and Museums(PHLM). Dave followed the websites instructions and pretty soon there were 45 people looking forward to the first PHLM meeting and 167 people in the group at large. At the first meeting Dave gave a small presentation, people introduced themselves and two members brought toys to show off and share. At the end of the meeting Dave asked for feedback and received a few suggestion.
A few days after their first meeting I met with Dave, he was worried, What do 167 people want? How would he sustain the interest of the group? Membership was growing daily but after talking to a few participants at the last event he was unsure what anyone wanted or expected from PHLM.
167+ people with diverse expectations.
No established goals.
Members self select into and out of the group. There is no active recruitment other the their meet-up listing.
Dave has one month before the next event
What activities can PHLM organize to satisfy member expectations?
The solution was quite simple, give people what they want and they will come back for more. The hard part was to determine what they wanted. If we did a survey and asked it would take more time then Dave had so asking them was out. The next best solution was to go look at the types of things members of PHLM did. Where better to look then in their Meet-up profiles.
We use various software tools to visit the meet-up website, go into each PHLM members profile and copy their list of groups they belong to. The next step would be to see what people did in these groups that Dave might reproduce. The data showed over 1700 distinct groups, too many to investigate individually. The list was centralized in a database and ranked based of the number of members who shared links to these other groups. Figure 1
shows the distribution of the groups based on number of shared PHLM members. The data follows a well known phenomena on how people organize information called Zipf's law. Zipf's law has been used since just prior to 1950 and the corpus of literature that has been developed around his work is mature and respected. Based on academic research utilizing Zipf's law certain assumptions can be made. In this case two constrains can be argued, 1) the most popular items were most important for predicting the majorities interests and secondly only the fist handful of records needed to be investigated before the return they gave became negligible.
At this point the first few most popular groups were viewed again and their event histories were scrutinized and recorded.
An eight page report outlining the purpose, methods and findings. Additionally several recommendations were made including two templates to base an event on.The templates blended the types of activities that were most common, a time line that recommended a later start for those commuting and examples of several groups event announcements to illustrate effective messaging illustrate effective messaging.
It was also recommended that participants be given feedback cards before they leave that investigate all the elements of the event, timing, activity, networking, etc.
Dave reported that the next event followed our template and he felt both confident and satisfied with the outcome.