The impact of project characteristic on the ICO performance

Размещено на http://www.allbest.ru//

Размещено на http://www.allbest.ru//

INTRODUCTION

project performance regression

In the modern economy, the introduction of technological innovations has an increasing influence on the nature of conducting a successful business, including changing the structure of the financial component of the entrepreneurial activity. Traditional sources of financing, such as obtaining bank loans, issuing bonds or placing shares on the stock exchange are replaced by decentralized and more effective mechanisms of fundraising: borrowing without intermediaries (peer-to-peer loans), crowdfunding and, completely new phenomenon, the issuance of crypto-currencies (ICO).

Based on the "blockchain" technology, the ICO mechanism allows small companies and start-ups to collect funds almost from the zero-stage of the project development, having only an attractive idea or business plan. Despite the high risks of investing in cryptocurrencies, the ICO market in 2017 exceeded the amount of venture financing in conventional currencies, and the capitalization of the market itself is already comparable to that of many industrial and technological giants and is approaching the size of the GDP of some developed countries.

The listed circumstances determine the relevance and importance of the topic of this work.

The purpose of the study is to identify factors that affect the success of the ICO. To achieve this goal, the following main tasks have been accomplished:

1. The basic concepts of ICO are covered.

2. The mechanism of ICO implementation has been considered and the state and development of the ICO market have been studied.

3. The analysis of the factors influencing the success of the ICO was carried out.

4. An econometric model is constructed to predict the success of the ICO.

The object of research in the work is the market initial coin offering. The subject of the study are the factors that affect the attraction of funds through the sale of tokens.

In this work, we used such analysis methods as correlation analysis, analysis of means, variance analysis, the creation of multiple regression models.

This work consists of an introduction, two chapters, and a conclusion. In the introduction, the urgency of the chosen topic is substantiated, the object and subject of the research are formulated.

In the first chapter, the theoretical basis for the analysis of the success of Initial Coin Offering - the basic concepts, the state of the market and the ICO algorithm, were considered.

In the second chapter, a statistical analysis of ICO success was carried out based on a sample of 243 projects sequentially monitored and listed on the website ICOdrops.com from August 2017 till the end of February 2018, the fundraising for which was completed at the time of data collection.

Besides, in order to complete the data, we performed the sentiment analysis of social network mentions, details of which will be explained further in Chapter 2.

Lastly, the main conclusions on the results of the study are formulated.



CHAPTER 1: THEORETIC BASICS OF THE ANALYSIS OF ICO SUCCESS

1.1 ICO's key definitions

Speaking of the ICO (Initial Coin Offering) first we have to define such things as cryptocurrency and blockchain technology.

History of the cryptocurrency development started with the Bitcoin which was made in 2009 by an unknown person or group of people who used the name, Satoshi Nakamoto. For the provision of its functionality and system's security cryptographic methods are used. All information regarding the transactions between addresses (wallets) within the system is available for the public access. Bitcoins' success led to the mass emergence of other cryptocurrencies which later was called "Altcoins" (basically meaning "alternative coins").

Thus, the definition of cryptocurrency can be the following: it is the digital currency which is created through the internet and stored on the private virtual wallets. When creating this currency, a special cryptographic code-cipher is used, which consists of a certain consecutive hashing of transactions and a digital signature. One unit of such cryptocurrency is called coin. Such coin is in fact only encrypted data that cannot be forged.

In contrast to crypto-currencies, traditional money (cash and non-cash) is usually called by the crypto-enthusiasts a fiat currency (from Latin "fiat" - “Let it be done”).

The functioning of bitcoins and other cryptocurrency is based on the blockchain technology. A blockchain is a digitized, decentralized, public ledger of all cryptocurrency transactions. Constantly growing as `completed' blocks (the most recent transactions) are recorded and added to it in chronological order, it allows market participants to keep track of digital currency transactions without central recordkeeping. Each node (a computer connected to the network) gets a copy of the blockchain, which is downloaded automatically https://www.investopedia.com/terms/b/blockchain.asp.

The schematic mechanism for implementing transactions in the blockchain is shown in Fig. 1.1.

Fig. 1.1. Algorithm for the functioning of blockchain technology

First, this technology was implemented as a component of the Bitcoin cryptocurrency, where the blockchain plays the role of the main register of all operations with bitcoins. Security of the blockchain is provided by a set of special cryptography programs. The database is distributed among a large number of servers and is autonomously provided by thousands of powerful computers included in the mining system.

Mining is the act of supporting a distributed database and creating new blocks. Miners process transactions and provide network security with specialized equipment mainly based on graphics cards (GPU) in exchange receive new bitcoins or another cryptocurrency https://bitcoinmagazine.com/guides/what-bitcoin-mining.

With the growth of the distributed network, the complexity of the calculations that the miner must do in order to confirm the validity of the block is also growing. This leads to the need to use more and more powerful equipment, as well as to increase energy consumption. That is the reason why another method of confirming transactions was developed - the algorithm for the confirmation of the user's stake (PoS - Proof of Stake). The idea of this algorithm is the following: the probability of creating a new unit and receiving a corresponding reward is proportional to the user's stake owned within in the system of all coins MM Pryanikov, AV Chugunov. Blockchain as the Communication Basis for the Digital Economy Development: Advantages and Problems // International Journal of Open Information Technologies. 2017. №6..

It is possible to distinguish the main characteristics of the blockchain technology:

Decentralization;

The openness of all data;

Cryptographic protection of information;

The inability to change the data that already entered into the system.

The blockchain can be divided into the following types: open, closed and combined. An open blockchain system is a system in which there is no restriction on the reading of blocks (i.e. any user of the network has access to all data), and the supervisory authority is completely absent.

There is also a closed blockchain: network, where there is a controlling authority, and access to data is only available to the organizations that created the network. It should be noted that with this approach, the very essence of the blockchain system as an open distributed registry is ignored. This generates a network vulnerability to hacker attacks.

Combined blockchain system is a type of technology that allows to take the best from both types and minimize their shortcomings. This blockchain contains both an element of the distributed registry and a controlling authority (which often required by the government). Comparative characteristics of these types of blockchain are given in Table. 1.1.



Table. 1.1.

Comparative characteristics of blockchain typesMM Pryanikov, AV Chugunov. Blockchain as the Communication Basis for the Digital Economy Development: Advantages and Problems// International Journal of Open Information Technologies. 2017. №6.

Characteristics	Blockchain type
	Open	Closed	Combined
Identification	Absent	Identification if the network members	Identification within the network
Access to the participation in network	No restriction on participation	Access to participation in the network of a narrow circle of participants	Access to participation in the network is limited by rules (for example, a network member can only view its transactions)
Status of validators	Status of the process is not assigned to the participants	Status of validators is assigned to certain counterparties	Status of validators is assigned to certain counterparties
Presence of supervisory authority	Supervisory authority is absent	There is a supervisory authority	There is a supervisory authority

One of the most promising areas of using blockchain-technologies today is to attract investments by issuing crypto-currencies which are called ICO (the initial offering of coins).

This mechanism allows start-ups and small companies to raise funds almost at the zero stage of project development, having only a business plan or just an attractive idea.

The essence of ICO is that the financing of a new project is due to the sale of future cryptocurrency (tokens). ICO is conducted both for crypto-currencies created on the basis of own blockchain and based on the existing algorithm such as Ethereum.

Tokens are digital coupons - replacement of coins which the based on blockchain network of some particular coin of any coins, to implement the task of lending and monetization of an additional crypto-currency service https://masterthecrypto.com/differences-between-cryptocurrency-coins-and-tokens/.

Tokens produced in the ICO process differ in their characteristics and functions. There is no clear classification of tokens for today, however, their main types can be distinguished depending on the functions they perform.

The most common kind of tokens is application tokens. The main function of this token is to grant access to a certain network or application. Typically, these tokens are analogous to cryptocurrency. Investors invest in them in the expectation that application tokens will grow in price when the issuing project will become successful.

As ICO is becoming an increasingly popular way to attract investment, there have been many start-ups that have an indirect relationship to the cryptocurrency. In this case, tokens (or token-certificates) give the investor the right to receive a certain asset (product, service). The disadvantage of such tokens is that ownership of them does not bring profit, since the token-certificate costs as much as the asset, which it is provided.

The second kind of tokens are share-tokens, which are in some way are analogs of real securities. Holders of such tokens are entitled to receive income (dividends) from the possession of tokens. Dividends from tokens may look like part of the projected revenue or a certain percentage of transactions on the Web R. Akst. 7 sekretov ICO. Ili pot u storonu tokenseila/R. Alst. - «Izdatel'skie reshenia», 2018. .

The advantages of owning a share-tokens for investors are the same as the ownership of ordinary shares. If the company that issued the tokens is successful, then the cost of the tokens will also grow, providing the owner with additional revenue.

The third type is credit token, i.e. tokens, which are issued in the course of the ICO as debt obligations to investors. The advantage of investing in such tokens is a high-interest rate (up to 100% per year), and a disadvantage is a high risk and the lack of regulation.

It should be noted that the classification of tokens is not legally prescribed and is rather vague. Often, tokens can perform functions of several types at once.

A useful tool for ICO participants is a smart contract. The main goal of smart contracts is to guarantee a secure exchange of assets between unreliable agents. In an ICO, a smart contract can ensure that in the event of a project failure, investors will get their money back, and when the developers achieve their goals, the collected funds will automatically drop into their account.

Smart contracts have three main properties - autonomy, self-sufficiency, and decentralization. Autonomy means that after the contract is stated, there is no need for further interaction with the initiator. The self-sufficiency of the contract secures the mobilization of resources and assumes that contracts are able to raise funds, provide services or issue securities, and spend them on the necessary resources, for example, processing power or storage. Smart contracts are decentralized, that is, they do not focus on one central server, but are distributed among the network, where they are independently executed, M. Swan. Blockchain: Blueprint for a New Economy: Melanie Swan / M. Swan. - O'REILLY, 2015 (First edition) - p.16

.

Like any program, a smart contract has a number of drawbacks:

1) the complexity of self-compilation of smart contracts (you need to use a special programming language);

2) dependence on the human factor (when writing a smart contract, there may be an error in the code that leads to its incorrect performance);

3) insufficient flexibility (the data entered into the blockchain cannot be changed).

Despite all the problems described, most experts agree that smart contracts are very promising for the development of the crypto-currency market. The platforms that provide secure tools for the implementation of smart contracts are Waves, Stratis, NEO, RootStock, Propy, as well as their ancestor, Ethereum.



1.2 Mechanism of raising funds through ICO

ICO or Initial Coin Offering -is the activity to attract capital to the project, based on the use of blockchain technologies. The funds are attracted to the project by selling tokens or coins that are emitted for a specific project that is sold directly to early investors with the discount and later usually placed on exchanges for free circulation on the market. It received its name by analogy with the IPO (Initial Public Offering).

The procedure for conducting an IPO and ICO is quite similar, but there are a number of important differences. First of all, it is worth noting that an IPO can only be held by an established company that has been successfully operating for a long time. Moreover, the company that carries out the IPO must take on a number of obligations:

* Ensuring stability of growth of financial indicators and increase of market share;

* Asset size and / or annual turnover of more than $ 10 million;

* Availability of an audit of the US standard and GAAP format statement;

* Carrying out activities with respect to investors' interests;

* There should be qualified personnel, including top managers who have a vision of the company's strategic development S. V. Kondratova, M.V. Umrihina. IPO kak istochnik finansirovaniya deyatel'nosti kompanii. // Nacional'naya associacia uchenykh (NAU), 2015. - №7, P. 109-113.

ICO, in contrast, is conducted primarily by start-ups, i.e. investors invest money in an idea that has not yet been realized (or is at the initial stage of implementation).

The second important difference is the legal regulation of IPO and ICO procedures. The IPO is conducted in accordance with the requirements of the legislation as well as exchange rules of the country in which the company operates and issues its shares. As for this moment, conducting of the ICO is not clearly regulated by the law of any jurisdiction.

It should also be noted that when conducting an IPO, investors buy a stock that gives them the right to own share in the statutory fund of the issuing company. ICO participants invest in crypto-currency funds in exchange for tokens, which have certain functions or promised profits as a share of income within the system developed and defined by the founders. This means that tokens do not give the right to own part of the share/property of the issuing company.

The differences between the IPO and ICO procedures are grouped in tab. 1.2.

Table 1.2.

Differences between the ICO procedure and the IPOBased on https://www.forbes.com/sites/rogeraitken/2017/01/06/investment-guide-to-crypto-coin-offerings-rating-blockchain-startups/

Characteristics	ICO	IPO
Investment object	Tokens	Shares
Costs of conducting	Minimal, from $10 000 mainly spent on marketing campaign	Minimum worth of conducting an IPO is around $200 000
Who conducts	Startups	Working and usually profitable business
Protection of the interests of investors	Not protected at the state level	Protected at the state level and can be sued for breach of obligations
Legal Form	Everything is mostly based on investors' confidence	Common form of ownership
Dividends	Dividends are absent	Dividends exist in the conventional form
Trade and Exchange	Trading tokens on crypto-currency exchanges that support this type of token	Trading on stock exchanges

Let's consider in more detail the procedure of conduction an IPO and ICO. The sequence of the company's IPO is shown in Fig. 1.2.

Fig. 1.2. Stages of the IPO procedure https://www.mergersandinquisitions.com/initial-public-offering-process-ipo/

The ICO procedure as a whole is similar to the sequence of stages of the primary placement of shares, it also includes the preliminary, preparatory and main stages, as well as the completion of the transaction. However, this procedure is much easier due to the lack of the need to comply with the legislative requirements and requirements of the stock exchange.

The preliminary stage of ICO includes the formulation of an idea of the project and the development of a business plan. Both are stated in the so-called white paper, which is the main document of the project. Moreover, at this stage, a team for the implementation of the project is formed.

The structure of the white paper is not regulated by anyone, however, most of them include the following items:

1. Background - information about the idea itself, the reason for its appearance, its full comprehensive description.

2. Introduction - a description of the scope of applicability of the idea, a description of the market, general provisions for the monetization of the project.

3. Market Analysis - the results of marketing research, confirming the prospects for the product to enter the market.

4. Introduction to product - a description of the future product.

5. Product specific section description of the technical component of the project.

6. Technical (description of technical aspects) - contains all the technical aspects associated with the production of the product (service) and the stages of implementation.

7. Usability - the practical value of the product being developed, its advantages for the investor.

8. Road Map - achievements and future plans on the timeline.

9. Financial Projections - distribution and use of investments, financial plan.

10. Team, advisors, experts, auditors and other project participants.

11. Terms - terms of use and other legal aspects of the project.

Thus, white paper is analogous to the prospectus of securities issued in the IPO.

Moreover, at this stage, the ICO organizers usually create a project website containing the following information:

* A brief description of the ICO and presentation of the product;

* The "roadmap", i.e. a list of planned activities with a breakdown by time;

* Information about the team;

* Documents (public offer, white paper);

* Project partners;

* The subscription form;

* Contact information (e-mail address, links to it on social networks, etc.);

* Investor's personal account.

Usually, the ICO site is made in the form of a single page (landing page). It performs several important functions at once: provides and advertise key information about the project and company and used as a platform for primary token sale (in the most cases, the sale of tokens is carried out through the investor's personal account on the site).

The preparatory stage of the ICO is called pre-ICO or ITO (initial token offering). This stage is necessary for the primary raising of funds and is actually crowdfunding (a method of collective financing, based on voluntary contributions).

The fundraising at the stage of pre-ICO is carried out by selling tokens to interested investors with various bonuses or discounts in gratitude for the early investment in the project. This procedure is called "crowdsale".

Also, at the pre-ICO stage, the date of the project's release to the ICO is announced, meetings with potential investors are held, the project information is posted at specialized forums and other methods to familiarize the participants of the crypto community with the new project is used. In general, this stage is similar to the "road show", i.e. meetings with investors, which are held during the IPO.

It should be noted that the preparatory stage of the ICO may not happen if the project has an investor who is ready to invest in the ICO or an authoritative leader who is promoting the idea of the project.

The third, main stage of the ICO is the stage of directly raising the investments for the project. At this stage, work is also being done with potential investors. PR-company of the project can be implemented in the following ways:

1) Project presentations at various conferences and personal meetings with potential investors;

2) Placing the project on specialized websites, for example, forums devoted to crypto-currencies, and ICO-trackers;

3) A marketing campaign in social networks, instant messengers, and search engines.

One of the tools of the project promotion is a Bounty Campaign. It is a reward program for helping to promote the ICO. Since participation in the Bounty Campaign is paid with tokens of the future project, it allows the founders to reduce the budget of the advertising campaign.

The most common types of work that ICO organizers are willing to pay tokens for are:

The Promotion of ICO in Social Networks - posts, reposts, comments, subscriptions, etc.;

* Comments on specialized forums on the web, for example, BitcoinTalk;

* Translation of information about the project into the other languages;

* Posting articles about ICO in thematic blogs and other resources;

* The creation and distribution of videos, podcasts etc.

The preparatory stage for the ICO usually lasts from a one to six months, and ICO itself takes several weeks.

1.3 History of development and the current state of the ICO market

The history of the ICO market began in 2012 when the software developer J.R. Willett published a "white paper", in which he proposed to create a programmable "money layer", a superstructure over the existing blockchain of the bitcoin. His proposed protocol ("Mastercoin”) was aimed to allow anyone to create their own cryptocurrency, on basis of bitcoin. In order to get the required financing for this development, in July 2013, the Mastercoin fund conducted an ICO: five hundred people who wanted to help the project transferred to its e-mail address 5 thousand Bitcoins which was about $ 500 thousand at that period's exchange rate, receiving in return digital tokens “UpToken”. The largest ICO for the entire period of the market was Ethereum, which in early 2014 collected more than $ 18 million https://www.rbc.ru/magazine/2017/11/59e62b5d9a7947de527907d4.

The dynamics of the number of ICOs conducted and the funds collected are shown in Table. 1.3.

Table 1.3.

The dynamics of the number of ICO

and the size of the raised investments Calculated using data from https://stats.digrate.com/

Indicators	2014	2015	2016	2017	2018 (January-April)
Attracted funds, $ mln.	29,23	7,364	123,186	6626,092	5928,383
Rate of growth, %	-	25,2%	1672,8%	5378,9%	89,5%
Number of ICOs conducted	5	5	50	886	648
Rate of growth, %	-	100,0%	1000,0%	1772,0%	73,1%
Average amount of investments into one project, $ mln.	5,846	1,473	2,464	7,479	9,149

From the given table it is clear that in 2014-2015 there was practically no development of the market, but since 2016 a boom started. The number of ICOs was 50 (10 times more than in the previous year), and the volume of collected investments - 123.186 million dollars (16.7 times more than that indicator in the previous year). In 2017, the growth rate of the ICO market was even higher. Obviously, in 2018, the market growth rate will also be high. Already on the basis of the results of four months (January-April), the amount of funds attracted is around 89.5% of the level of 2017, and the number of ICOs held reached 73.1% of that number in 2017.

It is also obvious that the average amount of investment per project is gradually increasing. In 2015, it was $ 1.473 million, and by 2018 it reached $ 9.149 million.

Graphically, the dynamics of the ICO market volume is shown in Fig. 1.3.

Fig. 1.3. ICO Market Dynamics

Rapid growth in market volumes does not mean that all projects that have done ICO are successful. A significant number of projects fail to conduct a successful ICO and collect all the planned amount of investment (some does not attract anything at all). It should also be noted that a successful ICO does not mean the success of the project itself. Many start-ups attract investment without having a product (prototype or working version), and after the completion of the ICO, the idea of a project often remains unrealized. According to the «RBC» (rbc.ru) study, of the 100 largest projects of 2017, only 11 projects had a working product or service, another 5 startups had a product or service with a partial functionality https://www.rbc.ru/magazine/2017/11/59e62b5d9a7947de527907d4.

Therefore, when analyzing the ICO market, it is advisable to consider not only the success of the project's ICO but also the success of the project as a whole. The share of unsuccessful ICOs and the share of "dead" projects in the total number of ICOs conducted is calculated in Table. 1.4.



Table. 1.4.

Analysis of the ICO success Calculated using data from https://stats.digrate.com/

Indicators	2014	2015	2016	2017	2018 (january-april)
Total number of ICO conducted	5	5	50	886	648
including:
successful ICOswhere company operates until today	5	4	33	319	348
In % of the total	100,0%	80,0%	66,0%	36,0%	53,7%
unsuccessful ICOs	0	0	1	306	287
In % of the total	0,0%	0,0%	2,0%	34,5%	44,3%
"dead" ICOs	0	1	16	261	13
In % of the total	0,0%	20,0%	32,0%	29,5%	2,0%
% of successful ICOs	0,0%	25,0%	48,5%	81,8%	3,7%
Total amount of attracted investments, $ mln.	29,230	7,364	123,186	6626,092	5928,383
including investments attracted by the "dead" ICOs	0	0,740	20,492	954,779	119,378
In % of the total	0,0%	10,0%	16,6%	14,4%	2,0%

It can be seen from the table that the rate of ICO success decreases annually. If initially, all ICOs were successful, in 2016 only 66% of the total are successful, and in 2017 only 36%. Increasing the share of successful ICOs in early 2018 is a positive trend.

"Dead" ICOs are projects that have development problems after the successful completion of the ICO fundraising. Signs of "dead" projects:

1) the project site does not work;

2) project tokens did not appear on stock exchanges within 4 months after the end of the ICO;

3) investors have sued the organizers of the ICO.

It can be seen from the table that the share of "dead" projects in the total number of ICOs is quite high (the results of 2017 - 29.5%), and their share in the number of successful ICOs (and "dead" ICOs are formally successful, as they succeeded to collect the necessary amount of investment) - is huge. In 2017, the share of "dead" ICO was 81.8% of the total number of successful ICO. This means that more than 80% of all start-ups that raised funds through ICO turned out to be fraudulent. However, if you look at the amount of attracted investments, the picture is not so pessimistic - the amount of investment in "dead" projects amounted to 14.4% of all funds collected through ICO.

It is also worth to mention that according to the results of 4 months of 2018, only 2% of the ICO's were "dead." This indicates that investors are gradually beginning to understand the situation, to be cautious and are keen to learn more about the project before investing in it.

The first projects that attract investments through ICO were, in most cases, projects related to blockchain technology and the development of crypto-currencies. However, other projects are gradually entering the ICO market. The sectoral structure of the ICO market (according to ICO-tracker www.coin-schedule.com) in 2017-2018. is provided in Table 1.5.

Table 1.5.

Sectoral structure of the ICO market Calculated using data from www.coinschedule.com

Category	Funds raised, $ mln.	In % of total:	Changes in the structure, %
	2017 г.	2018. (January-April)	2017 г.	2018 г. (January-April)
Infrastructure	1002,500	225,978	25,8%	3,5%	-22,4%
Finance	564,621	883,314	14,6%	13,5%	-1,0%
Investment and trading	386,323	700,728	10,0%	10,7%	0,8%
Communications	322,565	1819,954	8,3%	27,9%	19,5%
Payment systems	291,22	281,64	7,5%	4,3%	-3,2%
Databases	286,223	94,089	7,4%	1,4%	-5,9%
Pharmaceuticals and health	225,326	91,528	5,8%	1,4%	-4,4%
Games and Entertainment	133,418	438,817	3,4%	6,7%	3,3%
Other industries	667,822	1996,460	17,2%	30,6%	13,4%
Total:	3880,018	6532,508	100,0%	100,0%	-

It can be seen from the table that the main share in the volume of collected investments in 2017 was the category "infrastructure". This category includes projects that improve and develop blockchain technology and new crypto-currencies. From the calculations, we see that in 2018 the share of these projects in the total amount of investments decreased to 3.5%. The second most important place in the volume of investments is the "Finance" category. It includes projects that develop platforms for money lending, micro investments, insurance, exchange of crypto-currencies into real money, etc.

A fairly large share in the volume of funds attracted is given by the category "Investments and Trading" (Figure 1.4). This category includes mainly various platforms for investment in crypto-currencies and trading on crypto-exchanges.

Fig. 1.4. Sectoral structure of the ICO market in 2018

The weight of the category "Communications" significantly increased in 2018. This category includes projects based on the technology of blockchain mobile networks, social networks and other types of communication.

Also, it worth to mention that in the category "Other industries" there is a sector of the economy which has a rather large share in the total volume of the market - "Commerce and Advertising". The share of this sector in 2018 increased to 5.8% (in 2017 it was only 1.9%).

Thus, from the comparison, we see that the structure of the ICO market is gradually changing. Infrastructure projects such as the development of new crypto-currencies are replaced by projects that try to apply blockchain technology in other sectors of the economy.



CHAPTER 2. ANALYSIS OF FACTORS THAT INFLUENCE ON THE SUCCESS OF ICO

2.1 Correlation analysis of the factors influence on the success of the ICO

The main tool for assessing the relationship between variables is the correlation analysis, in particular, the calculation of the correlation coefficients. Two variables correlate positively with each other if there is a direct, unidirectional relationship between them. With a unidirectional relationship, small values of one variable correspond to small values of another variable and large values to large values. Two variables correlate with each other negatively, if there is an opposite direction between them, i.e., small values of one variable correspond to large values of the other variable and vice versa. Values of correlation coefficients are always in the range from -1 to +1.

The presence and nature of the correlation relationship can be determined from the scatter plot. Fig. 2.1 shows the scatter plot for the variables "ICO Success Rate" and "Token Price". The diagram does not show a correlation between the variables. Note that ICO success rate can take values above 1 (equivalent to 100%) meaning that project collected more funds than the goal it set.

To evaluate the correlation relationship between the indicators in the SPSS software, 3 correlation coefficients are used - the Pearson linear correlation coefficient and the Spearman's and Kendall's Tau rank correlation coefficients.

Pearson's correlation coefficient is used most often, but it can only be applied when the distribution of variables is normal or insignificantly different from normal. Therefore, before carrying out the correlation analysis, it is expedient to test the hypothesis of the normality of the distribution of the variable "ICO Success Rate".

Fig. 2.1. Scattering diagram for variables

"The success rate of ICO" and "The price of the token"

The normality of the distribution is tested using the histogram of the distribution (Figure 2.2) and the Kolmogorov-Smirnov (K-S) test.



Fig. 2.2. The histogram of the distribution of the variable "ICO Success Rate"

It can be seen from the figure that the histogram of the distribution of the variable is very different from the normal distribution curve. We formulate the null hypothesis in the following way: the distribution of the variable "ICO Success Rate" corresponds to the normal one. To test this hypothesis, the Kolmogorov-Smirnov criterion is calculated (Figure 2.3).

Fig. 2.3. Calculation of the criteria for the normal distribution

The P-value for the calculated statistics is 0.000, i.e. they are below the accepted significance level (б = 0.05). This means that the null hypothesis of the normality of the distribution was rejected. Thus, to evaluate the correlation relationship between the variable "The success rate of ICO" and other characteristics, nonparametric criteria-Spearman's and Kendall's Tau correlation coefficients should be used.

Consider the hypotheses of a correlation between the success of the ICO and the following variables:

* price of the token, $;

* The number of team members;

* Project mentions in Social Networks

In order to introduce a measure for the project mentions in Social Networks, we made a Python script using the Sentiment Analysis package and Twitter API. Only Twitter mentions were loaded as Twitter is the main channel for communication for most of blockchain-product based companies and its API allowed to load last 100 mentions for each project which in most cases was an accurate measure for the social opinion on the project at the moment of data collection. So, the variables used here are:

* Number of positive mentions on Twitter;

* Number of negative mentions on Twitter;

* Number of neutral mentions on Twitter.

Descriptive statistics for the variable "Price token" are shown in Table. 2.1. The sample included 234 observations, the minimum price was $ 0.001, and the maximum was $ 29.1. The typical value of the price of the token is $ 0.846. The standard deviation means that on average the price of the token in the sample deviates from the average value by $ 2,528.

Table 2.1.

Descriptive statistics for the variable "Price token"

	Number of observations	Minimum	Maximum	Average	Standard deviation
Price of the token, USD	234	,001	29,100	,84617	2,528281
N of valid (full)	234



Correlation coefficients between the variables "ICO Success Rate" and "Token Price" are shown in Fig. 2.4.

Fig. 2.4. Spearman's and Kendall's Tau correlation coefficients

for variables "ICO Success Rate" and "Token Price"

The Kendall Tau coefficient is -0.182, the Spearman coefficient is -0.249. Both mean that there is an inverse correlation relationship; The lower the price of the token, the more successful ICO. Since the values of the coefficients are below 0.3 (in absolute values), the correlation is rather weak. Nevertheless, it is statistically significant at a significance level of б = 0.01. Thus, the price of the token is one of the factors affecting the success of the ICO.

To estimate the correlation between the success of the ICO and project mentions in social networks, we calculate the Spearman's and Kendall's Tau coefficients for these variables. The results of the calculations are given in Table. 2.2.



Table 2.2.

Correlation coefficients for the ICO Success Rate variableand the number of mentions on Twitter

Indicators	Positive Twitter mentions	Negative Twitter mentions	Neutral Twitter в mentions
Kendall's coefficient	0,120	-0,041	-0,006
Relevance	0,009	0,386	0,890
Spearman's coefficient	0,170	-0,056	-0,008
Significance	0,008	0,388	0,895

The table shows that there is a correlation between the positive mentions of the project on Twitter and the success of the ICO. The Kendall's Tau coefficient for this pair of variables is 0.12, and Spearman's is 0.17. This indicates a direct correlation of weak force. The significance of the calculated coefficients is close to zero, i.е. they are statistically significant at the level б = 0.01.

The significance of the correlation coefficients for the remaining two variables is much higher. This means that the null hypothesis on the equality of the correlation coefficient to zero was confirmed (that is, the calculated coefficients are not statistically significant). Hence, we can conclude that the success of the ICO is affected only by positive feedback in social networks (Twitter in our case).

Consider the variable "Number of team members" (Table 2.3).

Table 2.3.

Descriptive statistics for the variable "Number of team members"

	Number of observations	Minimum	Maximum	Average	Standard deviation
Number of team members	217	2	120	13,72	13,937
N valid (full)	217

The number of observations for the variable "Number of team members" was 217. The minimum number of team members for the project is 2 people, the maximum - 120 people. The average value of the variable according to the sample is 13.72, i.е. a typical number of team members is 14 people.

The values of the correlation coefficients for the variables "ICO Success Rate" and "Number of Team Members" are shown in Fig. 2.5.

Fig. 2.5. Results of correlation analysis for variables

"The success rate of ICO" and "Number of team members"

Kendall's Tau correlation coefficient was 0.065, and Spearman's was 0.089. Since the coefficients are below 0.10, this indicates that there is no correlation relationship. The significance of the coefficients is much higher than 0.05, which confirms the null hypothesis that the correlation coefficients are equal to zero.

Thus, the number of team members does not affect the success of the ICO.

The correlation dependence between nominal variables can be estimated using contingency tables. To do this, we recode the variable "ICO Success Rate" into a binomial variable with two values:

1) the success rate is less than 1 - ICO failure;

2) the success rate of 1 or more (i.e., more funds were collected than planned) - the success of the ICO.

We construct the crosstabulation table for the variables " ICO Result" and "Presence of the Prototype" (Table 2.4).

Table 2.4

Crosstabulation table for variables "ICO Result" and "Presence of a prototype"

	Is there a Prototype?	Total
	No	Yes
ICO Result	ICO failure	Frequency	20	29	49
		% in ICO Result	40,8%	59,2%	100,0%
	ICO success	Frequency	4	24	28
		% in ICO Result	14,3%	85,7%	100,0%
Total	Frequency	24	53	77
	% in ICO Result	31,2%	68,8%	100,0%

The table shows that the number of observations for the variable "Presence of a prototype" is only 77. This is because for many projects there is no data on the presence of a prototype. Among the failed ICO, 40.8% of the projects had a prototype, and 59.2% did not. In successful ICO, the share of projects with a prototype was 85.7%. Obviously, the variables are interrelated, and the presence of a prototype increases the probability of ICO's success.

This hypothesis can be verified using conjugacy criteria. To evaluate the relationship between nominal variables, we can use the Chi-square test, as well as the contingency coefficient and thecriteria for Phi and V Cramer. The above criteria do not give an answer to the question of the direction and strength of the relationship between the variables, they are used only to test the null hypothesis on the independence of variables. Since the Chi-square test can be used only if each cell contains at least 5 observations, we use other contingency criteria (Figure 2.6).

Fig. 2.6. Criteria for contingency for the variables "ICO Success"

and "The presence of a prototype"

The contingency coefficient is 0.266, V Cramer is 0.276, and Phiis 0.276. The significance of all three criteria is the same and is 0.016. Since this value is below the accepted significance level (0.05), the zero hypothesis on the independence of variables is rejected. Thus, there is a connection between the presence of the prototype and the success of the ICO.

Table 2.5 presents the table of contingencies between the variables "ICO Result" and "The Possibility of Paying with a Fiat Currency"

Table 2.5.

Contingency table for variables

"ICO Result" and "The Possibility of Paying with a Fiat Currency"

	Possibility to pay with fiat currency	Total
	No	Yes
ICO Result	ICO failure	Frequency	112	14	126
		% in ICO Result	88,9%	11,1%	100,0%
	ICO success	Frequency	106	10	116
		% in ICO Result	91,4%	8,6%	100,0%
Total	Frequency	218	24	242
	% in ICO Result	90,1%	9,9%	100,0%

It can be seen from the table that the number of observations for these variables was 242. The ability to pay with a fiat currency is absent in 90% of all reviewed projects. Approximately the same ratio is observed for groups in the context of the success of the ICO. In the group "The failure of ICO", the possibility to pay with the fiat currency is absent in 88.9% of the projects, and in the "Success ICO" group - 91.4%. Thus, the relationship between the variables is not revealed.

To test the hypothesis of the independence of variables, we perform the chi-square test (Figure 2.7).

Fig. 2.7. Testing the hypothesis on the independence of variables "ICO Success" and "The ability to pay with a fiat currency"

The Chi-square value is 0.419 at significance value of 0.517. Since the significance is much higher than 0.05, the null hypothesis on the independence of variables is accepted. This means that having the opportunity to pay with a fiat currency does not affect the success of the ICO.

2.2 Mean and variance analysis

The influence of qualitative factors on the effective variable is assessed using methods of comparing the means and variance analysis. These two methods solve similar problems, but the average comparison method is used to compare the mean values in two samples, and the variance analysis allows us to compare the mean values in several samples.

The method of comparing means is performed in following way: the data set is divided into two subgroups (samples) based on the qualitative variable, and the average value is calculated for both samples. Next, the null hypothesis about the equality of the mean of two samples is formulated, which is tested with the help of the Student's t-test. If the null hypothesis is confirmed, the difference in the average values for the samples is said to be insignificant. If the null hypothesis is rejected, it means that there is a statistically significant difference between the mean values ??for the samples, i.e. the qualitative factor underlying the division into subgroups affects the outcome variable. This procedure is called a t-test for two independent samples.

Let us check whether the presence of a prototype affects the success of the ICO (Table 2.6).

Table 2.6.

Selected average values of the ICO success rate depending on the presence of a prototype

Presence of a prototype	Number of observations	Average	Standard deviation	The standard error of the mean
ICO Success rate	No	24	0,2054	0,40925	0,08354
	Yes	53	0,7117	0,38949	0,05350

It can be seen from the table that the average ICO success rate for the second sample (projects having a prototype) is much higher than the first. This means that the presence of a prototype has an impact on the success of the ICO - the existing prototype product (service) increases the probability of successful fundraising. The calculation of Student's t-test is shown in Fig. 2.8.

Fig. 2.8. The result of the t-test for independent samples

In addition to calculating the t-test, the software program conducts Levene's test for the equality of variances. Since the significance of the F-statistics is 0.788, the hypothesis about the equality of variances is accepted. Therefore, the result of the t-test is looked at the first line of the table "Assuming the equality of variances". The t-statistics is -5.201 with a significance of 0.000, i.e. the null hypothesis about equality of means should be rejected. Thus, the t-test confirmed the assumption that the presence of a prototype affects the success of the ICO.

Also, using the t-test for independent samples, it is possible to assess the impact on the success of the ICO by the availability of the option to pay with the fiat currency and the number of currencies accepted. The average values of the ICO success rate for the samples, depending on the possibility to pay with fiat currency, are given in Table. 2.7.

Table 2.7

Selected average ICO success rate

depending on the possibility of paying with fiat currency

The possibility to pay with fiat currency	Number of observations	Average	Standard deviation	Standard error of the mean
ICO Success rate	No	218	0,6714	0,43641	0,02956
	Yes	24	0,6463	0,41606	0,08493

It can be seen from the table that the average value of the samples is almost the same - for the first sample it was 0.6714, and for the second sample, it was 0.6463. The result of the t-test is shown in Fig. 2.9.

Fig. 2.9. The result of the t-test for independent samples

in the context of the opportunity to pay with a currency

The significance of the F-statistics is 0.361, i.e. the hypothesis about the equality of variances is accepted. The t-statistics is 0.269, the significance is 0.788. This means that the null hypothesis about equality of means is confirmed. Thus, the conclusion drawn earlier that the possibility to pay with fiat currency does not affect the success of the ICO has been confirmed.

To assess the impact of the number of accepted currencies on the success of the ICO, we divide the aggregate into two samples - ICO, in which only 1 currency is accepted, and ICO, in which 2 or more currencies are accepted (Table 2.8).

Table 2.8

Selected average ICO success rate

depending on the number of currencies accepted

Number of currencies accepted	Number of observations	Average	Standard deviation	The standard error of the mean
ICO Success rate	>= 2	85	0,6362	0,44901	0,04870
	< 2	157	0,6866	0,42553	0,03396

It can be seen from the table that the average values for the samples differ insignificantly: for the first sample, the average value was 0.6362, and for the second sample it was 0.6866. The results of the t-test are shown in Fig. 2.10.

...