ALGORITHM UPDATES HISTORY

Our reasoning for our algorithm is clear as daylight. 99.9% of the review and rankings websites are tilted in favor of the sellers and not the buyers. At CrowdReviews.com, our sole mission is to build a transparent algorithm for the benefit of the buyers so that they can make smarter purchasing decisions.

The following elements were introduced to our algorithm to make sure our rankings become more beneficial with each iteration.

 

  • June 23, 2017 – CRAlgorithm Ver. 0.85b

    Numerous changes have been made to increase the importance of relevant and detailed reviews and to protect the system against those which have an interest in gaming the system. Other changes have been made to provide incentive to reviewers to consistently produce fair and detailed reviews of the software and services they have used. As a result, some adjustments have been made in point values to account for the new factors.

    • The algorithm now takes into consideration reviews that are left by people which have a proven history and identity, the type of information that is presented in a review which would indicate the reviewer has extensive experience with the service or software, reviews which may not be particularly fair to the service or software being reviewed based on an emotional appeal, and the level of originality expressed in the thoughts found within a review. Other changes have been made to account for the new factors.
    • Changes have been made in the weightage of different factors used to determine reviewer score to account for the inclusion of the industry component, giving more points for reviewers in industries where the service or software are more prominently used. Those which have an extensive history are given more value in terms of their impact as a reviewer if they have a history of doing business within a relevant industry compared to those which may have less experience.
    • Engagement Strength has also been adjusted with the addition of self-prescribed FAQs and quarterly profile updates ensuring the service and software providers are investing in their engagement with their reviewers, clients, and potential clients. Service and software providers are also incentivized to take part in a video interview with a member of the CrowdReviews.com team for the benefit of both providers and helping potential buyers learn more about those behind the services and software found on the website.
    • We have implemented a number of different potential penalties for companies which may engage in activities which would be seen as detrimental to the fairness of the position of other services and software on the platform. A penalty has also been introduced to take action against actors which may attempt to exploit the platform or otherwise undermine the performance and accuracy of the platform.
    • New factors have been added to both the “How Do We Rank” page as well as the forecasting tool offered to all service and software providers, allowing for them to test out the changes made to the algorithm.
  • June 15, 2016 – CRAlgorithm Ver. 0.69b

    Significant updates have been incorporated to our algorithm in this update. We have merged Reviewer Strength and Review Strength, introduced Sentiment Strength, upgraded the ability for the algorithm to recognize the quality of the reviewer, further strengthened the ability to recognize insightful reviews, introduced the requirement that a service vendor must have a physical location for it to be eligible for rankings in that country, and recalibrated our scoring system across all variables.

    • We have merged Reviewer Strength metric into our Review Strength metric. We felt the Reviewer had just as much to do with the review itself. We did keep all the elements of how we measure the quality of the Reviewer within our algorithm for the benefit of our users. Additionally, this enables us to introduce a new major element to our algorithm for the benefit of our buyers.
    • Sentiment Strength. We now make use of tools that use machine learning models and multiple signals (including web and social) to determine the public sentiment of the companies in our database. Instead of trying to rebuild these tools, which have been developed in some cases over a decade, we incorporated these tools in our algorithm and gave it 20% value to the overall score given to a company. This helps our visitors significantly in that it doesn’t only incorporate reviews left on our website, but also how all users on the web collectively feel about companies on our rankings.
    • LinkedIn Review Adjustments. As companies try to game the rankings, we continue to calibrate our algorithm to the benefit of our visitors and buyers of such products and services. Our algorithm now takes a closer look at the LinkedIn Reviewer themselves and the types of reviews they contribute. One LinkedIn reviewer is no longer the same as another LinkedIn reviewer. Buyers will now get rankings that are influenced by legitimate reviewers rather than LinkedIn reviewers created to influence rankings.
    • Insightful Reviews adjustments have been made to give more weightage to those reviews that are deemed as more useful. Buyers expect reviews to be helpful, although reviews such as “Company XYZ sucks” can be considered valuable, we have now incorporated additional weightage to reviews that are more explanatory and have concrete details.
    • Localization. This new attribute applies only to service vendors on our website and not software products. We now require that a service vendor have a physical location in the country they wish to be eligible for. This enables buyers to connect with companies that are more likely to be able to assist them, especially for services that are location or region dependent.
    • Scoring System. Some value adjustments have been made to give slightly less importance towards the Verification Status that should be expected from all vendors and more weight towards Sentiment Strength and Engagement Strength. Less importance has also been placed on Profile Strength since it is relatively simple for a company to complete their profile.
  • February 8, 2016 – CRAlgorithm Ver. 0.43b

    We introduced Insightful Reviews and Content Engagement as additional factors in our algorithm generating even more valuable rankings for the benefit of those considering to purchase products and services.

    • Insightful Reviews has been introduced to give more weightage to those reviews that are deemed as more useful or helpful by our visitors.
    • Content Engagement has been introduced to give more value to companies in our algorithm that engage and provide content of value to our visitors, such that, they respond to questions that have been asked to them, upload whitepapers or content of value on their company profile page, and those that engage and contribute insightful content relevant to their industry on the CrowdReviews.com Blog.
  • December 18, 2015 – CRAlgorithm Ver. 0.28b

    We introduced Engagement Strength and Reviewer Strength as two new major areas of analysis. To ensure authenticity of reviews, we incorporated a LinkedIn account login process before a review is left. For the benefit of our visitors, we set a requirement for companies to get reviews on their profile pages before their profile could be considered complete.

    • Engagement Strength replaced Reputation Strength, however, we kept the elements of measuring the Reputation Strength active and upgraded the algorithm to set the foundation for measuring other areas of engagement than just analyzing the responses to negative reviews.
    • Reviewer Strength is now incorporated in the algorithm to evaluate the quality of the reviewer. Weight distribution differs based on if the reviewer is a customer, an employee, or a non-disclosed party.
    • LinkedIn Reviews enables reviewers to leave a review tied to their identity. With so many anonymous reviews on other websites, we give far more value to reviews that have an identity attached to it.
    • Minimum Review Requirement. We renamed “profile completeness” to “profile strength” and included a minimum number of reviews obtained for a profile to be considered complete. At least 5 reviews are required for a company to receive full credit for profile strength.
  • October 1, 2015 – CRAlgorithm Ver. 0.1b

    We introduced Review Strength, Profile Completeness, Reputation Strength, and Verified Status as the major components in our first iteration of the algorithm.

    • Review Strength focuses on the average star ratings and the age of the reviews.
    • Profile Completeness gives an added advantage to companies that have taken the time to complete their profile.
    • Reputation Strength reviews the number of negative reviews a company receives, the responses posted, and then further analyzes how many negative reviews are resolved.
    • Verified Status is granted to companies that have their accountants / attorneys verify their key facts to ensure accuracy.