First of all I want to thank you for sharing this information because it was really information packed. It really got me thinking
I looked at the ClassDojo graph you showed and saw that the tipping point was 3 teachers per school, where a teacher refers the other 2.
I couldn't understand how you were able to know that a teacher referred another teacher. I mean it could be that a teacher from a school joined ClassDojo but didn't refer the other teacher from that same school, and that other teacher heard about ClassDojo on her own. Did you use attribution surveys to know that? Or did you use other ways.
Also don't you think the WOM coefficient will not be accurate if we rely on attribution surveys, given, as you said attribution survey results are often unreliable?
Thanks for your questions. Let me take a crack at answering:
“how do we know a teacher referred another teacher”.
—isolation: there was no other marketing or promotion of the product. Zero paid ads to muck this up. No multiplayer mode for the product so no incentive to invite someone else to improve your experience.
—did lots of interviews with teachers. The pattern was very clear from that: teacher to teacher or teacher to staff meeting (1 to many)
“WoM coefficient and attribution surveys”
—WoM coefficient was not done on ClassDojo data for clarification. It calculated on a bunch of gaming companies as well as other non-gaming verticals and had high R^2 when done correctly—which provided a ton of value for growth models and growth planning.
—WoM coefficient does not use attribution surveys—it uses regular marketing attribution data (so direct + non blog page organic + undefined traffic = word of mouth)
It is not easy to find content laser-focused on WOM. Congrats and thanks for sharing Yousuf
Thanks Jose!
This is an awful lot of expertise to just throw out there for free. Thank you!
First of all I want to thank you for sharing this information because it was really information packed. It really got me thinking
I looked at the ClassDojo graph you showed and saw that the tipping point was 3 teachers per school, where a teacher refers the other 2.
I couldn't understand how you were able to know that a teacher referred another teacher. I mean it could be that a teacher from a school joined ClassDojo but didn't refer the other teacher from that same school, and that other teacher heard about ClassDojo on her own. Did you use attribution surveys to know that? Or did you use other ways.
Also don't you think the WOM coefficient will not be accurate if we rely on attribution surveys, given, as you said attribution survey results are often unreliable?
Thanks alot man, appreciate it.
Thanks for your questions. Let me take a crack at answering:
“how do we know a teacher referred another teacher”.
—isolation: there was no other marketing or promotion of the product. Zero paid ads to muck this up. No multiplayer mode for the product so no incentive to invite someone else to improve your experience.
—did lots of interviews with teachers. The pattern was very clear from that: teacher to teacher or teacher to staff meeting (1 to many)
“WoM coefficient and attribution surveys”
—WoM coefficient was not done on ClassDojo data for clarification. It calculated on a bunch of gaming companies as well as other non-gaming verticals and had high R^2 when done correctly—which provided a ton of value for growth models and growth planning.
—WoM coefficient does not use attribution surveys—it uses regular marketing attribution data (so direct + non blog page organic + undefined traffic = word of mouth)
Hope that helps.
Hi Yousuf, I really appreciate your in-depth reply.
Yeah makes sense.
Your blog is really really insightful. Please keep on writing blog posts.
Thank you.