A New Experiment: Automating and Monetizing a Gift Idea Blog

After stumbling upon the ability to call Open AI’s GPT3 through an API in R, I had a multitude of ideas for useful outputs. After considering the effort each of these ideas would require and the potential to go awry in a harmful way, I settled on starting with a blog for gift ideas at ai-gift.com . While there are still ways this could go awry, I feel it would take minimal human oversight relative to alternative ideas. I will hold off on sharing the alternative ideas, although I’m sure the ideas could be independently arrived at without much effort for those familiar with the functionality of GPT3.

Reasons Why I Like This Idea and Associated Goals

I’ll break the reasons for liking this idea up into 2 parts, the practical and the technical/experimental. Then I will discuss the two goals of increasing traffic/monetization at minimal effort and actually providing value to users.

Practical Reasons For Settling on This Idea

  • High volume of potential content (holidays, types of people, groups, etc.)
  • Coming up with prompts for the content is low effort and GPT3 davinci is pretty good at answering the prompts.
  • Can copy/paste output into gift idea blog but would rather move towards a more efficient upload
  • Can create a bunch of blogs on a weekend, and have them on release schedule spread out over time
  • By curtailing prompts, should be able to avoid potential harmful output (unlike, say, asking for a political opinion or financial advice)

Technical and Experimental Reasons for This Idea

  • While the content creation is relatively straight forward (produce lists of gift ideas, do a tl;dr on a product description, synthesize a group of reviews, etc.), the growing of the audience for this content is the challenge.
  • How can SEO increase traffic?
  • How can social media be automated in sync with posting to circulate updates and promote articles?
  • How can some form of marketing increase traffic (leaving cards on pinboards when traveling, enlisting friends in different GEO’s to leave cards on pinboards)
  • Practice at defining KPI’s to guide strategy
  • Should allow for the creation of anonymized data to try more advanced analytic approaches to increasing KPI’s (This is the interesting part to me)

Goal 1: Increasing Traffic/Monetization at Minimal Effort

  • Track time and track views/monetization from ads/affiliate links. The goal for all of the KPI’s derived from these data is to increase the number representing a higher return on time. Constraints include 2 hours of time investment per week.
    • KPI 1 = views in a week / hours put in that week
    • KPI 2 = cumulative views / cumulative time invested
    • KPI 3 = weekly monetization / weekly time
    • KPI 4 = cumulative monetization / cumulative time invested

Goal 2: Providing Value To Users

  • Keep track of time spent on pages and once links and ad data are available track those as well as time invested. The goal again is to increase the value of these KPI’s which I think are indicative that some value is being provided.
    • KPI 5 = Average page view time that week / hours that week
    • KPI 6 = cumulative page view time / cumulative hours spent
    • KPI 7 = Ads clicked in week / hours that week
    • KPI 8 = Cumulative ads clicked / cumulative hours spent

I’ll probably aim to give a monthly update here on the metrics, data gathered, strategy, and changes.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: