Rambles

Decision Process for Key Metrics for DenseLayers

By January 5, 2020 January 17th, 2020 One Comment

 

The first version of DenseLayers is very close to completion (it’s past the deadline of Dec 31, which makes me disappointed but at the same time I’m glad I’m making a lot of progress and doing it right.

Need to decide what kind of metrics I should be monitoring as the founder, during the crucial initial stage. Using this post to think out loud from first principles and put my ideas on paper, hoping it will help me make sense.

What I want to accomplish

I want to grow the website into a product that people enjoy using, and can imagine themselves using at least on a weekly basis.

This means a few things:

1. They see DL as a place where discussions are of high quality
2. They find the website easy to use

If the above were true, users would respond with by:

1. Visiting the website often
2. Reading their paper in detail
3. Reading other people’s comments posts
4. Adding more of their own comments posts
5. Signing up with an account! (duh)
6. Being nice to each other

I should stick to using the word “post” as opposed to “comments” from now on, since comments are a different feature that will not be in V1.

Possible metrics derived from each of these items:

  1. Overall traffic to different pages
  2. Traffic to different papers
  3. Number of logins per week?
  4. Recurrent visits (monthly active/weekly active/daily active)
  5. % of all signed up users who log in/visit per week
  6. Change in traffic from week to week
  7. How many posts people are writing
  8. Number of posts per paper
  9. Number of users who have ever written a post on a paper
  10. % of people who have visited each paper, who wrote a post
  11. % of visitors per week, who wrote a post
  12. % of contributors who show spammy behaviour

Now – I want to be really careful in choosing my key metrics here. The metrics I choose will trickle down into every other decision I make. They will define the whole strategy. And focusing on the wrong metric can jeopardize long-term sustainability of the DenseLayers community. I could argue that I should definitely be *looking* at all of these metrics (or build in a way to calculate them), but I should make my decisions based on only 1-3 metrics that reflect long-term health and growth, and trump everything else.

My way of choosing the right metrics here would be to go back to the overarching things I want to accomplish. For each metric, I’ll ask the following questions:

1) Does this directly mean users enjoy using DenseLayers?
2) Does this directly mean DenseLayers is a place for high quality discussions?

For the first version of the site, these two are extremely important questions. So let’s go through the list of metrics again:

Overall traffic to different pages
Traffic to different papers
Number of logins per week?
Recurrent visits (monthly active/weekly active/daily active)
% of all signed up users who log in/visit per week
Change in traffic from week to week
How many posts people are writing
Number of posts per paper
Number of users who have ever written a post on a paper
% of people who have visited each paper, who wrote a post
% of visitors per week, who wrote a post
% of contributors who show spammy behaviour

 

The three metrics marked in bold should be the main metrics. Weekly active users is an obvious number, it shows people who enjoy coming to DenseLayers regularly for whatever reason. Number of posts per paper is important because having a paper on DenseLayers is useless if people aren’t discussing it. And then, I want as many people as possible to write posts and not just read them, so the % of weekly visitors who write a post captures that.

I will be calculating and looking at some of the other metrics though. So I can refer to them if I find anomalies in the data.

 

 

One Comment

Leave a Reply

Designed by

best down free | web phu nu so | toc dep 2017