Project Everest

Adopted Experiment

[FEB 19] Facebook (Channels) - FarmEd II Fiji

Marcin Poloczek
Marcin Poloczek | 11 months ago | in FarmEd - Fiji

Assumption: Farmers are willing to interact with FarmEd through Facebook


Lean Phase: Channels


Time Box: 3 weeks


Key Metrics:

Number of page likes from Fijians over the experiment period


Success Point:

70 Page Likes


Green Light:

Proceed with Facebook as an advertising channel. Investigate viability of Facebook as a distribution channel over the long term.


Orange Light Range:

20-69 Page Likes


Orange Light:

Continue testing Facebook to determine effectiveness over the longer term. Continue to increase engagement by sharing page within groups, inviting existing contacts to like the page through Hubspot details, or via other stakeholders.


Failure Point:

0-19 Page Likes


Red Light:

Reconsider the content and frequency of existing posts, and investigate other means of promoting the Facebook page (e.g. sponsored posts).


Experiment Build:

The purpose of this experiment is to demonstrate whether it is viable to establish a large enough audience in order to test Facebook as a distribution channel in future experiments. The experiment will be reviewed at the end of each week to redefine metrics. If green light is achieved, engagements will be assessed as a measure of willingness to buy, as opposed to likes.


Goal: 6 posts on page throughout the experiment and weekly posts in buy and sell groups



  • Create alias Facebook account to admin and share the page
  • Create Facebook page for the FarmEd application
  • In the description of the page include all relevant information on the app and PEV



  • Share page into existing groups
  • Spread word through WOM when meeting with stakeholders
  • Create content to share into page twice a week
  • Reply to any messages within 4hrs and redirect sales to FarmEd I



  • If green light reached early, assess key metrics to evaluate Facebook as a distribution channel (i.e. number of engagements - direct messages, shares and comments, as a proportion of total reach)
  • Record any feedback received to further develop the application

Brodie Leeson 11 months ago

This is a fantastic experiment, really valuable for gathering data on channels.
Some points and questions:
- Could be valuable to post a link to the page in this post, non essential but after having a look at the page myself it seems to be really well run so could be beneficial for readers to have easy access to it.

- Are you Filtering out likes by PEV people in your assessment of success metrics?

- How do the interactions contribute to success metric, are they simply a means to accomplish the likes or could they carry their own weight on success. Do the posts on Buy and sell pages always direct people to the page or can they sometimes hold their own value, like providing insight into something through a comment. Could be beneficial to see the structure being used to post into buy and sell groups so that it could be continued or refined moving past February.

Reply 0

Laura Taylor 11 months ago

Hi Brodie! Some good questions!
We are definitely filtering out likes by PEV people. We will only consider likes by Fijians for reaching our 70 like success metric.

For February we decided to focus our efforts on gaining page likes and so all posts made are to engage current followers or to direct other people to liking our Facebook page. The reason for this is that we felt any experiments that compare number of post likes/comments/shares/etc. vs. number of people reached would be too subjective a measure if we only had a small number of targeted viewers. These kind of metrics we do believe could be valuable for a future experiment.

We have not conducted any direct interactions in buy and sell groups on other people's posts. Our goal for channels is to reach lots of people with minimal effort from us so interacting on individual posts didn't seem to add value to our experiment.

Reply 0

Lucy Preiss 10 months ago

Status label added: Experiment adopted

Reply 0