Uncategorized

A study of 1200 ICOs and more than 100 factors that influence fundraising

Client’s ICO investment problem

Our client was from a South African company that provides online cross-border payments with new technologies, including Blockchain settlement.

They started an ICO project and faced the challenge of choosing the best marketing activities to invest in. They turned to Singularex for a predictive analysis.

The main goal of the work was to understand how much money that an ICO could raise and what channels the firm could promote to obtain at least $10M.

Machine learning to predict ICO returns

Team

Two analysts conducted this research for two weeks.

Instruments and methods

The predictive analysis was performed with the help of neural networks, decision trees, and clustering.

Process Stages

  • Collecting data from cryptocurrency platforms and other sources
  • Selecting the most significant parameters
  • Use of different Machine Learning algorithms
  • Analysis of the results
  • Predicting the ICO return

What factors most influence ICO fundraising

We managed to collect data on 1200 ICO projects and studied more than 100 of their characteristics. After analyzing this dataset, we defined the four most crucial factors that influence on ICO projects:

Alexa ranking

The higher the Alexa Rank, the more funds the ICO project will raise.

Repository stars on Github

The more repository stars the ICO project has, the more funds are raised.

Percentage of token distribution

95% of ICOs distribute between 43% and 79% of tokens in an ICO.

A further increase of distributed tokens by percentage doesn’t lead to an increase in the amount of collected funds.

Telegram members

The more Telegram members the ICO community has, the more funds will probably be raised.

The results of the predictive analysis

With the help of clustering, all the ICO projects we examined were divided into 6 groups.

After we studied all the previous ICOs and their characteristics, analyzed the current state of the client’s ICO sand used Machine Learning models to predict its return.

The results of our analysis were the following:

  • If Alexa rank increases by 1 step, raised funds will increase by $3M
  • If Github Stars increase by 1, raised funds will increase by $1M
  • If Token distribution increases by 1%, raised funds will decrease by $65 k
  • If the Telegram Members Score increases by 1 step, raised funds will increase by $1M

According to the research data, we predicted that the ICO would raise $10M if they took into account the results of our analysis for marketing activities.

How this study helped the client to raise $12M

The results of our research formed the basis of the client’s marketing activities for their ICO promotion. The distribution of tokens began one month later than it was planned. Ultimately, about $12M was raised through our client’s ICO.

That means that our predictive analysis was reliable.

Get a free consultation from our expert

Categories: Uncategorized