Experience with Blockchain/Cryptocurrency and Feelings About Brand Overall
ANOVA Lower Bound Upper Bound
No 427 3.958 .7680 .0372 3.885 4.031 1.0 5.0
Yes 146 4.171 .6471 .0536 4.065 4.277 3.0 5.0
Total 573 4.012 .7443 .0311 3.951 4.073 1.0 5.0
Table 28: ANOVA Descriptives
The results of an Analysis of Variance (ANOVA) show that those who have invested in blockchain assets before are significantly more likely to feel positive about the brand overall (p = .003). The difference is relatively small, but those who have invested reported feeling slightly more positive (mean = 4.17) than those who have not (mean = 3.96).
73 7.6 Regression
Several regression analyses were conducted to determine the relationships between variables, including likelihood to use SKILLITY. This variable was collected on a scale from 0 (not likely) to 2 (likely)
Gender and Likelihood of Using SKILLITY
Coefficientsa
Table 29: Regression Coefficients, Gender and usage
Gender is significant in predicting whether someone can see themselves using SKILLITY (p
= .001), although the variance explained is small (R-square = .018). Males are more likely to indicate that they would use SKILLITY than females.
Age Range and Likelihood of Using SKILLITY
Analysis of Variance
Table 30: Regression Coefficients, Age Range and usage
74 Age Range is significant in predicting whether someone can see themselves using SKILLITY (p
= .018), such that younger individuals tend to be more likely to see themselves using SKILLITY;
however, the effect is very small (-.025 points for each “unit” of age range).
Independent Worker and Likelihood of Using SKILLITY
Analysis of Variance
Table 31: Regression Coefficients, Freelancing status and usage
Whether someone is a freelancer, self-employed, or small business owner (or other independent worker) is significant in predicting whether they can see themselves using SKILLITY or not (p
< .001). Those who are independent workers are more likely to report that they would consider using SKILLITY.
Experience with Cryptocurrency and Likelihood of Using HEART Tokens
Past Investment in Blockchain Assets
Likelihood of… Overall Model Variable
F-value p-value R-square t-value p-value
Purchasing HEART Tokens 48.74 <.001 .081 6.98 <.001
Exchanging HEART Tokens 76.06 <.001 .119 8.72 <.001
Accepting HEART Tokens as Payment 58.97 <.001 .095 7.68 <.001
Making Payments in HEART Tokens 49.29 <.001 .081 7.02 <.001
Table 32: Experience with Blockchain influencing the likelihood of token usage
75 Those who have invested in blockchain assets in the past are more likely to purchase HEART tokens than those who have not (p < .001). They are also more likely to exchange HEART tokens (p < .001), accept HEART tokens as payment (p < .001), and make payments in HEART tokens (p < .001).
Past Use of Cryptocurrency
Likelihood of… Overall Model Variable
F-value p-value R-square t-value p-value
Purchasing HEART Tokens 30.71 <.001 .052 5.54 <.001
Exchanging HEART Tokens 54.36 <.001 .087 7.37 <.001
Accepting HEART Tokens as Payment 35.54 <.001 .058 5.96 <.001
Making Payments in HEART Tokens 51.91 <.001 .084 7.20 <.001
Table 33: Experience with Cryptocurrency influencing the likelihood of token usage
Just as those with experience investing in blockchain assets are more likely to use HEART tokens than those without, the same is true of those with experience using cryptocurrency in the past (p
< .001).
7.7 Logit Model
Logistic regression (Logit Model) is used to model the relationship between the binary dependent variable and the independent variables. This analysis is undertaken in order to create a model (a regression formula) that can predict with at least some degree of accuracy which outcome is likely to follow a certain set of variables. In this case, we are attempting to determine whether someone is likely to use SKILLITY or not based on several demographic variables, experience with cryptocurrency and freelancing platforms, and attitudes about SKILLITY. The Likelihood to use SKILLITY is measured against all alternative solutions to offer or book services. These alternative methods include competitor platforms, local service companies or analogue ways to book freelancers (in person) and were not further specified in the questionnaire and analysis. For this
76 analysis, the likelihood of using SKILLITY was coded as 0 and 1, with “No” and “Maybe” coded as 0 and “Yes” coded as 1.
Predicting Likelihood to Use SKILLITY
Model Summary
Step -2 Log likelihood
Cox & Snell R Square
Nagelkerke R Square
1 592.888a .197 .271
a. Estimation terminated at iteration number 5 because parameter estimates changed by less than .001.
Table 34: Logit Model Summary Hosmer and Lemeshow Test
Step Chi-square df Sig.
1 14.933 8 .060
Table 35: Hosmer and Lemeshow Test Classification Tablea
Table 36: Logit Model Classification Table Variables in the Equation
B S.E. Wald df Sig. Exp(B)
Step 1a Independent Worker .312 .213 2.144 1 .143 1.366
EverInvestedBlockchain .498 .248 4.035 1 .045 1.645
AppealofCoopOwnership .672 .123 29.617 1 .000 1.957
EverUsedBuyingorSelling 1.098 .229 23.050 1 .000 2.999
ComfortAcceptingCrypto .295 .089 11.069 1 .001 1.342
Constant -5.152 .580 78.892 1 .000 .006
Variable(s) entered on step 1: Independent Worker, EverInvestedBlockchain, EverUsedBuyingorSelling, AppealofCoopOwnership, ComfortAcceptingCrypto.
Table 37: Logit Model Variables
The best logistic regression model to predict likelihood to use SKILLITY uses the respondent’s answers on the following questions as independent variables:
• Whether the respondent is an independent worker or not
• Whether the respondent has ever invested in blockchain currency before
77
• Whether the respondent has ever used a freelancing platform as a buyer or seller before
• The appeal of cooperative ownership
• The comfort level of the respondent in accepting cryptocurrency as payment
All five are significant (p < .05) and the model predicts likelihood to use SKILLITY correctly 73.6% of the time. This means that we can be quite certain that these five variables have an impact on reported likelihood to use SKILLITY.
The Hosmer and Lemeshow Test is not significant at the p < .05 level, indicating that the model’s coefficients are not equal to zero (i.e., that the variables do have some impact on likelihood to use SKILLITY).
The coefficients show that for each response, the higher the value, the more likely the respondent is to use SKILLITY. For example, since the coefficient is positive for the experience with blockchain variable (EverInvestedBlockchain), this indicates that those who have previously invested in blockchain (variable value = 1) are more likely to use SKILLITY than those who have not (variable value = 0). The p-value for this coefficient is .045, making it a statistically significant finding.
However, although the model is significant, it explains a relatively small portion of the variance (Cox & Snell R square = .197; Nagelkerke R square = .271), indicating that this model explains an estimated 20% of the variance in likelihood to use SKILLITY. This suggests that there are other, unknown factors that are also important in predicting likelihood to use SKILLITY.
78 7.8 Cluster Analysis
Hierarchical Cluster Analysis
First, a hierarchical cluster analysis was conducted. However, the results were not entirely conclusive (see Appendix A for results). To further investigate the clusters in this data, several k-means cluster analyses were conducted and compared.
K-Means Cluster Analysis
Seven separate k-means cluster analyses were conducted, and the results compared. The analyses specified a different number of clusters each time, determined to be reasonable solutions to the problem of customer segmentation:
To determine which solution best fit the data, the sum of the mean error and the sum of the distance from the center of the cluster were calculated and graphed. The researcher used the “elbow method”
to determine which solution was best.
Figure 4: Sum of the Distance from the Cluster Center by Number of Clusters 650
79 This chart was helpful, but not conclusive. It seemed that the elbow occurred around the 6- or 7-cluster solutions but this was not definitive. Next, the researcher graphed the sum of the mean error to determine again where the “elbow” appeared.
Figure 5: Sum of the Mean Error by Number of Clusters
This chart was more conclusive on where the elbow appeared. Around the 6-cluster mark, the points levelled out to around 0.7. This suggested that the 6-cluster solution may be the solution that best fit the data.
K-Means Cluster Analysis – 6 Clusters
Once narrowed in on the 6-cluster solution, the next step was to make sure the 6 clusters were practical representations of customer segments.
The table below was put together to compare the clusters on all relevant variables. It shows some distinct differences between the clusters, on everything from demographics to experience with freelancing and cryptocurrency to the perceived value of SKILLITY’s benefits and likelihood to use SKILLITY and/or HEART tokens.
0 0.5 1 1.5 2 2.5
0 1 2 3 4 5 6 7 8 9
Sum of the Mean Error
80
6-Cluster Solution
Cluster Name Enthusiasts Sellers,
not Buyers Dual Users Accepting
Crypto Sceptics Dual Users Not
Accepting Crypto Non-Users
Ease of Finding Local Sellers (Higher number = greater difficulty)
3.1 3 3.1 3 2.9 3.1
Worry about Security of
Transactions
3.6 3.5 3.7 2.8 3.4 3.6
Worry about Quality of Work
as a Purchaser
3.9 3.8 3.8 3.6 4.2 3.9
81
Likelihood of Paying in HEART
Tokens
4.1 2.6 3.3 1.5 1.6 3
How Problematic are Bad Sellers? (Lower number = less of a problem)
82
Due to Bad Experiences Worried about Freelancing and
Feature 3 Mostly Male
and Early 30s Mostly Male
and Late 20s Mostly Male and Late
20s Half Male and
83 Services each cluster would
consider purchasing
Service 1 Tutoring or
teaching (49%) Handyman or
Repairs (44%) Cleaning (50%) Handyman or
Repairs (44%) Handyman or Repairs
(57%) Cleaning (43%)
Service 2 Cleaning (47%) Photography
(30%) Handyman or Repairs
(50%) Cleaning (40%) Cleaning (56%) Handyman or
Teaching (26%) Special Delivery and
Transportation (36%) Tutoring or Teaching (30%) Table 38: k-means 6-Cluster Overview
For a full listing of the services that each cluster would be interested in purchasing, see the table below.
Cluster Caretaking Cleaning Custom Art Handyman
or Repairs Live
Music Photography
Special Delivery and
Transportation Tour
Guide Translation Tutoring or
Teaching Other
84 The results of the cluster analysis indicated that there are six distinct groups of respondents:
1. Enthusiasts: those who are highly open to trying SKILLITY. They are generally experienced freelancers and many of them have used cryptocurrency and/or invested in blockchain assets before. They are optimistic about SKILLITY and do notbelieve that bad buyers or sellers are much of a problem in freelancing platforms. This group is the most likely of all groups to report intentions to use HEART tokens. They are mostly male (68%) and the largest age group is 30 – 34. According to the in chapter 5.1 presented McKinsey segmentation of independent workers, this cluster can best be described as “Free agents”. These respondents are experienced and professional freelancers. Professionality and economic independence can be expected, due to their high degree of blockchain experience and cryptocurrency usage.
2. Sellers: those who are interested in selling on SKILLITY but are unlikely to make purchases.
They are also experienced freelancers and at least somewhat familiar with cryptocurrency.
They are interested in hosting more skills or services than the other groups (mean = 3.2 skills) and they find low transaction fees to be the most valuable feature (mean = 4.2). They are also mostly male (58%) and are most often in their late 20s. According to the in chapter 5.1
presented McKinsey segmentation of independent workers, this cluster can best be described as “Reluctants”. These very experienced freelancers highly value monetary features and social security. Moreover, they rather prefer to be paid in local currency then in
cryptocurrency. Hence their self-employment might be out of necessity, but a primary income.
3. Dual Users Accepting Cryptocurrency: those who are interested in SKILLITY, but cautious due to previous experiences with freelancing platforms. They are also generally having some experience in freelancing and are somewhat familiar with cryptocurrency, but they are much more concerned about bad buyers and sellers than the first two groups. They
85 report that finding local sellers and service providers is not too difficult for them, and they worry about the security of their transactions on an online freelancing platform. Finally, they see ratings and reviews and social security as valuable features. According to the in chapter 5.1 presented McKinsey segmentation of independent workers, this cluster can best be described as “Casual earners”. These respondents would both offer and seek skills and mostly value nonfinancial features. Furthermore, they are willing to use cryptocurrency for transactions. It can be derived, that any possible income on the platform would be
supplemental and their preferred choice.
4. Sceptics: those who feel negatively or are doubtful about freelancing, freelancing platforms, and/or using cryptocurrency. The sceptic group is the least willing to try SKILLITY. They have little experience with freelancing or with cryptocurrency, they worry about the security of their transactions on an online freelancing platform, and they worry about the quality of the services they would purchase. They also find SKILLITY’s features less valuable than other groups. This group is roughly half male (51%) and the largest proportion of them are in their late 20s.
5. Dual Users Not Accepting Cryptocurrency: those who are interested in using SKILLITY but hesitant about the cryptocurrency aspect. Unlike the sceptics, this group is interested in trying SKILLITY but nervous about using cryptocurrency. They may have done some light freelancing before, but they generally have no experience using cryptocurrency. They struggle to find local providers and they find some of SKILLITY’s features very valuable, but they are very uncertain about using HEART tokens. They are also roughly half male (52%) and generally in their late 20s. According to the in chapter 5.1 presented McKinsey segmentation of independent workers, this cluster can best be described as “Financially
86 strapped”. These users might offer skills as supplemental income, because their main job is not paying enough for a living. Hence they are both offering and seeking skills on the platform but prefer to get paid in cash.
6. Non-Users: those who are somewhat interested in buying on SKILLITY but are unlikely to sell. This group may occasionally purchase services on SKILLITY but are uninterested in offering their skills or services. They have the least freelancing experience of any group (only 39% are independent workers and only 47% have any freelancing experience) and they report that SKILLITY’s features are not very valuable to them. Unsurprisingly, the mostly-buyer group believes that bad mostly-buyers are not much of a problem. Yet surprisingly they are very willing to use cryptocurrency and are also somewhat interested in
cooperative-ownership. Therefore, is could be assumed, that this group might include of investors, rather than users. They are the only group that is composed of more females (52%) than males (48%), but they are also generally in their late 20s.
7.9 Qualitative Analysis
At the end of the survey, two questions were posed to respondents on what, if anything, would make them hesitant to use SKILLITY and if they had any other comments, concerns, or suggestions about SKILLITY’s platform. These questions were worded as follows:
1. Is there anything about SKILLITY that makes you hesitant or uninterested in trying it?
2. Do you have any other comments, concerns, or suggestions regarding this freelancing platform?
Responses to this question were analysed by cluster membership to determine the biggest concerns and hesitancies of each group.
87 Hesitancy about Using SKILLITY
Cluster 1 (Enthusiasts) generally had few concerns about SKILLITY. Forty-seven out of the 69 respondents that provided an answer had no concerns or hesitancies about using the platform. Nine respondents had concerns related to the use of tokens/cryptocurrency, such as the variable value and desire to be paid in more established currency. Five respondents noted that their only concern is that SKILLITY is new and untested, meaning it does not have a track record yet. Finally, there were a few other concerns held by one or two respondents in this group, including security, ability to use SKILLITY to bring in income, availability in the respondent’s country, and trustworthiness of the other users.
Many respondents in Cluster 2 (Sellers) also had no concerns about SKILLITY; 21 out of 71 respondents indicated they had no major concerns or hesitancies about SKILLITY. However, the largest group reported concerns about being paid in HEART tokens or uncertainty about how it works (28 respondents). Ten respondents indicated that their only concern is the novelty of the platform, and that they would need to learn more before considering using it. Two respondents noted their dislike of the name “SKILLITY” and ten respondents reported other concerns, like transaction security, hidden charges, and the certification of quality by SKILLITY.
The respondents in Cluster 3 (Dual Users Accepting Cryptocurrency) have more concerns than those in the first two groups, although about half report no issues with the idea. Again, the biggest concern of this group is in the use of cryptocurrency as payment. Respondents worry about how it works, whether it will retain its value over time, etc. Twelve respondents noted that it is too new for them to take the risk yet. Three respondents were concerned about the pricing and charges for using SKILLITY and another three felt that SKILLITY does not offer anything over and above what other freelancing platforms offer. The rest of the responses (14) mention things like disliking the name or
88 the logo, concerns about legal issues and insurance, and whether it is committed to social justice and equality.
Respondents in Cluster 4 (Sceptics) were most concerned about the cryptocurrency aspect of the platform (31 respondents). They generally are not familiar with cryptocurrency or are worried about the instability of cryptocurrency. Only 17 out of 70 respondents noted that they had no concerns with the platform, and the rest were worried that the platform is too new for comfort, what would happen to users if it failed, what kinds of fees and charges they would see, and whether SKILLITY conducts any vetting or security checks of local buyers/providers.
True to their label, most respondents in Cluster 5 (Dual Users Not Accepting Cryptocurrency) who left a response indicated that the use of cryptocurrency makes them hesitant to use SKILLITY. Again, respondents were nervous about the instability and fluctuations in value, and in general did not have much understanding of how HEART tokens work. Six respondents were worried about investing in a new and untested concept, and the rest of the respondents in this group mentioned things like security, data and privacy, curiosity about the full-time benefits offered, and disliking the name.
Finally, Cluster 6 (Non-Users) was generally not very concerned or hesitant about using SKILLITY, although there were unlikely to mention selling on the platform themselves. Those who did have concerns mentioned cryptocurrency most often (22), wondering how it will survive in the changing climate, worrying about fluctuating value, and generally noting their lack of knowledge about cryptocurrency. Eight respondents mentioned that it is such a new service and that they would need more information on it before giving it a try. Finally, eight people had other concerns, such as worries about availability in the respondent’s location, the fees charged, and whether SKILLITY would take the side of the buyer or the seller in a dispute.
These findings underscored the fundamental difference between the six clusters.
89 Suggestions and positive Feedback
Cluster 1 (Enthusiasts) were predictably positive about SKILLITY in their comments. Forty of them had no comments and 13 left positive comments and well wishes. Two respondents suggested aggressive advertising, since the platform needs to host many options to be successful. Other comments included a question on competitive bidding, a suggestion on preventing scams, and concerns about the ability to accept non-cryptocurrency forms of payment, quality control, and the ease of use.
Respondents in Cluster 2 (Sellers) also generally had no comments (34) or positive things to say (12).
Other comments included encouragement to advertise on Facebook and Instagram, a comment on the name and logo, a suggestion to focus on the buyer’s perspective, and a general comment about needing to try it first before making any judgments.
One of the positive comments from Cluster 2 was:
“If what it says is true, it sounds like it could be the next new freelancing site. There is no freelancing site that I really like right now.”
Those in Cluster 3 (Dual Users Accepting Cryptocurrency) also had few comments overall (56), although 14 respondents wished the company luck as it moves forward. Three comments centered on the payment methods, suggesting that sellers should be able to choose from a range of payment options.
Five other comments noted concern about the newness of the platform, questions on when and where it will be available, and the fees charged, and a suggestion to advertise more. One respondent echoed
Five other comments noted concern about the newness of the platform, questions on when and where it will be available, and the fees charged, and a suggestion to advertise more. One respondent echoed