Fantastic Web 3 Advocates and Where to Find Them - Part 1 of 2

Part 1

1. Introduction
2. Data Specification

Part 2

3. Fantastic Web 3 Advocates and Where to Find Them

4. Key Takeaways and Concluding Remarks

Important Disclaimer

This content is provided for informational and educational purposes only, and should not be relied upon as legal, business, investment, or tax advice. Furthermore, references to any securities or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Factomind cannot be responsible for your use of the information provided in this content. Factomind has established, maintained, and enforced strict internal policies and procedures designed to identify and effectively manage conflicts of interest related to its business activities.

1. Introduction

A project kicked off and devs are building. The founders want people to know about their project, so they decide to tap L1 partners, VC investors, and retail advocates to help them showcase the projects—but out of thousands of advocates, whose doors should they knock on?

Founders are advised to partner with a Layer 1 Foundation that is highly supportive in building a community, yet there is no conclusive proof for them to verify the claim-to-be-supportives. Moreover, being supportive could very much digress from being actually helpful; its assistance could well be fruitless. All these questions are often left unanswered, or are answered in subjective forms.

We will try to provide a rule-of-thumb for choosing marketing partners, solely regarding the community power. What we WON’T provide: “A chain is superior to B chain, so choose A.” Instead, what we will provide: “A chain has been more supportive and more effective in a project’s Twitter growth than B chain, so choosing A is advisable if your priority resides in the Twitter community.”

In our previous paper, ‘The Hitchhiker’s Guide to Web3 Marketing’, we specifically listed out ready-to-use marketing tactics to be administered on Twitter, and emphasized on how it is important to message ‘legitimacy.’ Along the course, we also suggested making use of L1 and VC partners—or, third party agents— to convey the message.

As a follow-up for the guidance, we quantified the effect of advocacy shoutouts by various categories, and with this data, will try to provide a criteria for L1s, VCs, and retail advocates regarding choosing the right partners for community building.

In short, this paper will provide a guideline on which marketing partners should founders go for, when the founders are looking to improve upon their Twitter presence.

2. Data Specification

2.1. Data Collection

All our datasets in this paper were collected using the Twitter API. The scraped data comprises the following:

– Name (ex. Factomind)
– Handle (ex. @factomind)
– Location
– Description
– Follower
– Following
– Listed
– Posting
– Account Creation Date
– Verified
– Like
– Retweet (Reply)

We collected above data points for ~4,000 advocate handles and ~250 project handles. For advocate handles, we surveyed 600+ Twitter crypto lists and gathered respective list memberships.

From the above data points, we derived the following community metrics called interactions. For a detailed definition of interactions, please refer to our previous paper here.

From the above metrics, we derived derivative metrics as below:

D0 = The day before a Tweet
Influential Factor_D1 = The rate of growth in interaction on the day of a Tweet
Influential Factor_D2 = The rate of growth in interaction after a day from a Tweet
Interaction Change_D1 = The growth in interaction on the day of a Tweet
Interaction Change_D2 = The growth in interaction after a day from a Tweet

Our final advocate dataset would look like below.

For our data for Social Media volumes, we used data obtained from major data vendors and cross-checked with our internal database. No raw data has been used since we used derivative statistics instead. An example time series for our Social Media Volume data would look like below:

Table 1 – The Dataset Description of 3,802 Advocates

2.2. Research Methodology

Consideration 1 : All interactions to be counted

We used interactions, not organic interactions, for computing our derivative metrics (influential factor, interaction change) because in this paper, our focus is to quantify the impact of the shoutouts from advocates, rather than measuring the soundness of a project’s community itself.

Consideration 2 : Direct mentions only

Also, we only recorded those Tweets directly mentioning a project’s Twitter handle or containing a project’s ticker as ‘shoutouts.’ For example, if an advertising Tweet contains ‘@ethereum’ or ‘$ETH’ within their texts, it is taken as a shoutout, but if the Tweet only has ‘Ethereum,’ it is eliminated from the shoutout samples. The reasoning behind this decision is recorded in Appendix.

Consideration 3 : Echos only

Plus, the project itself has to Reply to or RT the exact shoutout Tweet, or the Tweet thread header containing the shoutout, for the shoutout to be counted in our analysis. This process was taken only to simplify data collection.

Consideration 4 : Identification of Twitter handles

Moreover, to account for non-official account shoutouts, we grouped Twitter handles together if expressly stated likewise in the description. For example, @elonmusk has an empty description on its profile, so it would be considered as a separate entity from @Tesla; however, @cz_binance has stated “CEO @binance” in its description, thus @cz_binance and @binance would be examined in tandem.

Figure 1 – How to Calculate Interaction Change_D1 and Influential Factor_D1

Influential factors and interaction changes are calculated based on the time of a shoutout Tweet. If a shoutout Tweet from an advocate happened on May 10th, the influential factor_D1 would measure the rate of change in interaction from May 9th to May 10th, the influential factor_D2 would measure the rate of change in interaction from May 9th to May 11th; the interaction change would measure the flat number increase. The influential factor_D1, would be the basis for our analysis as it most directly relates to the community response to shoutouts; other metrics, namely an influential factor_D2 and interaction changes_{D1, D2}, would be examined if we are looking at a large subset of the sample population, but will be ignored if we look at a small dataset to avoid excessive overfitting. To look into the methodical rationale behind choosing the influential factor_D1 as the basis and using moving average interactions not organic interactions to compute the factor while employing the latter metric to evaluate a project’s community, please refer to Appendix.