There are currently very few practical methods for assessing the reliability of employee’s references. This problem can be solved thanks to reputation systems based on blockchain and crowdsourcing – says Kartik Mandaville, CEO in SpringRole, in Bithub.pl’s #SundayInterview.
What’s your idea and why does job/recruitment market need SpringRole?
Kartik Mandaville: The job market is flooded with the unverified resumes. Reference checks are the way by which an employer checks the validity of the information that a potential employee has provided in his/her resume. The information provided varies and may include previous job experience, educational information, specific projects, and skills that the candidate possesses.
An estimated 53% of resumes contain some amount of factual inaccuracy. Thus, validation is very important in hiring a new candidate. If the documents or claims are not valid they can become a major concern for employers and can lead to rejection of the candidate or business risks down the line. Effective reference checking can prevent companies from making bad hires and weed out candidates that make false claims. Most reference checks involve checking attestations first made long ago, and these checks are repeated by many subsequent reviewers.
Reference checks and attestations are time consuming…
KM: And expensive. there are currently very few practical methods for assessing the quality or the reliability of such claims apart from manual verification or engagement of third-party verification services. Current attestation systems are also inadequate while dealing with forged or inaccurate credentials – this can involve a lot of middlemen and is susceptible to fraud. The fraud risk increases if the attestation is across national boundaries or when translation is required (e.g., university applications).
For example, an interviewee may claim to be a graduate of Stanford and extremely skilled at Python. These two claims require two entirely separate attestation processes. One, the interviewer must talk to the administration at Stanford to figure out if the candidate has been a student there. If so, during what period of time? What were their subjects and grades? Next, the interviewer must conduct a separate attestation to confirm the Python skill claim by checking certificates, talking to previous employers and colleagues, or administering a test or challenge – all of which is hectic and time-consuming.
And here comes the reputation system…
KM: Indeed. Reputation systems are programs that allow users to rate each other in online communities in order to build trust through reputation. Reputation systems are essential for any system where trust has to be established without personally knowing the person. These systems allow individuals to have a sense of reliability about a particular person.
What are the flaws of current reputation systems?
KM: Current reputation sources are flat and have little or no weight adjustment for the rater’s expertise. We believe that reputation technology needs to be weighted according to the skill of the person whose opinion is taken into account. That is, the impact of Alice’s attestation that Bob possesses a given skill should be based on the reputation and skill level of Alice for that skill.
And here comes SpringRole with solutions to these challanges. What are they?
KM: SpringRole is the platform where people can view, share and get attestations on their professional profile, thereby creating a verified resume that they can share and use. A user’s professional profile contains three parts:
– educational qualifications,
– work experience,
Each of these three will have their own flows to get attested and once they are verified, we write them to the blockchain. These claims, once verified, will not have to be re-verified in the future and can be relied upon. Claims related to educational and employment history will be verified by the respective organizations and can often be resolved with a True/False attestation. On the other hand, verifying the claim of being well-versed in a particular skillset is often harder. Other platforms have increasingly been incorporating tests and challenges to prove proficiency. While this does make it more objective, these exams can be gamed and must be maintained and standardized by a central authority. They also suffer from narrow scope – they are only able to capture the facets of skillsets that are objective in nature. Moreover, skillsets are not an exact science and there are varying degrees of proficiency in a particular skill.
We tackle this problem by crowdsourcing a user’s reputation on a particular skill. People in the user’s network endorse his/her proficiency in a skill, independently or at the user’s request. The number of people endorsing it and their individual scores in the particular skill that they are endorsing lets us compute a score for each user per skill and is a way to assess the skill level of a person. In appropriate situations, examinations are easily incorporated in this approach; but it can capture much more nuance in skill level.
Is crowdsourcing safer solution than current centralized systems?
KM: We believe that using our community is the best way to rate a person’s skills. Current endorsements systems like LinkedIn, Upwork have certain problems that make them ineffective in being a reliable tool for skill evaluation. One of the problems is that endorsements are free, a person is free to endorse as many people as they want in whatever categories they feel like. No weight is given to the expertise of the person that is endorsing. The second problem is that endorsements don’t need the other party to accept or reject the endorsement – this leads to spam and users getting irrelavant endorsements in their profile. SpringRole takes on these problems head on in the following manner:
1. Skill endorsements have a small cost that is attached to it.
2. The person who is being endorsed has the option to accept or reject an endorsement.
3. The weight of the person endorsing in taken into account in his / her endorsement.
To make the system fair and transparent, the cost of endorsement will depend on several factors like the skill scores of the endorser, endorsee and the rate of acceptance of endorsements.
To be continued.
Interviewed by Przemyslaw Cwik