
There’s a fascinating case study of internet-influenced professional reputation in the news right now, and we thought we’d offer some comment.
Often, Unvarnished’s approach to professional reputation is mistakenly compared to anonymous blog comments. Well, today we have an example where blog comments—a tool woefully under engineered for this purpose—were used for a rather messy discussion about professional reputation.
At bottom, this is an unfortunate episode that demonstrates the widespread demand for ways to surface and assess professional reputation information, the fact that the current tools people use for this are woefully inadequate, and why we think that Unvarnished is the better way. You can see why, below.
The Back Story

For those who are unfamiliar, a former Yahoo! engineering executive, Madhu Yarlagadda, was recruited by Skype as their Chief Development Officer. TechCrunch covered the move in a story, and that’s where things started to get very interesting.
Almost immediately after the post ran, a series of critical comments about Madhu started surfacing in the comment section of the article, accusing him of various unpleasantness, purportedly from former coworkers at Yahoo!. There were some positive ones as well, but the majority were negative, voluminous, and quite detailed.
Fast forward a month and Madhu has left Skype, as covered by the Ashlee Vance of the New York Times, for “personal reasons”. According to one of Vance’s sources, the comments “caught the attention of Skype executives” and it went from there.
Wrong tool for the job

Without trying to comment on the accuracy of the comments about Madhu, it’s exceedingly clear from this episode that a. there is a serious need for a way to surface this information, and that b. a comment thread on a TechCrunch article, or other blog, is certainly not the place to do it.
Below are the reasons, on a case by case basis, why blog comments are not suited for this sort of thing, and why we think how Unvarnished is set up is a much fairer, more professional alternative via our many safeguards.
3rd party verified identity v. robust anonymity – Not the same thing.
TC – Robust, unverified anonymity: On TechCrunch, which uses WordPress’ commenting system, commenters are truly anonymous. Their identities are not verified by a third party login (Facebook, Twitter, etc.), and the only thing that readers see are self-reported names.
Further, TechCrunch doesn’t have much more information about commenters than readers do, in that the only identifying features they see beyond what readers do is the commenter’s IP address.
The result is that it’s exceedingly easy for a single person to post as many comments as they want, using different pseudonyms. And if they wanted to be really sneaky about it, they could even make it hard for TechCrunch to identify them, by simply jumping between coffee shops.
The result is that bad actors can exploit weaknesses in the system, and crowd out the good faith commenters trying to surface valid information. And this reality makes it impossible for readers to distinguish between good information and bad, and to use that information for real decisions.
UV – Protected identity, but third-party verified: Contrast this to Unvarnished, where, while the identities of reviewers are protected, to allow candid, honest reviews to surface without fear of persecution, we rely on third-party verified identity, provided to us by Facebook Connect, to ensure that each reviewer is, indeed, a real person, and that a single person can only leave one review. You can read more about our identity verification efforts in this blog post.
Persistent Identity – Reviewers need reputations too

TC – No Persistent Identity: Related to the fact that TechCrunch does not require third-party identity to prove real “humanness”, is the fact that commenters do not have “persistent identity”. As a result, commenters have little incentive for good behavior, and little downside for abuse. If their IP address gets banned, they can just get a new one (not to mention their DSL / Cable provider likely gives them a new one every few days anyway). Good behavior cannot be rewarded, and bad behavior cannot be censured because commenters cannot establish their own reputations.
UV – Persistent Identity, Accountability and Incentives: Contrast this to Unvarnished, where reviewers have persistent identity to which their actions on the site are tied, over time. Consequently, good actors gain reputation over time, and are rewarded for it with enhanced reviewer reputation, which is then presented to readers to evaluate via a reviewer badging system.
Conversely, bad actors, whose reviews violate Community Guidelines, are reported as abusive, carry that behavior with them, and can lose their voice on the site as a result.
Because reviewers have reputation associated with their persistent identities, good behavior is rewarded, and bad behavior is censured.

Administrator Moderation – A firm hand on the tiller
TC – No Unified Moderation: TechCrunch takes pride in having initially unmoderated comments to enable vigorous, free-wheeling debate. Comments are eventually reviewed, but go live first. It makes lots of sense for most of their stories. And while they do report that they will remove comments in certain situations, they don’t have a delineated policy on what is and isn’t grounds for removal.
This lack of solid policy is actually addressed in this case, in Michael Arrington’s comments in a follow up blog post about the rationale for leaving up the anonymous, negative comments about Madhu:
“We spend a lot of time moderating comments and generally try to remove random negative stuff about individuals. But in this case the outpouring of negativity was detailed and overwhelming.”
UV – Community Guidelines-based Company Moderation: Unvarnished, on the other hand, engages in active moderation of reviews, reviewing each review that comes onto the site to ensure that it conforms to community guidelines. We detail our efforts to do this in this blog post.
For example, in the comments about Madhu made on TechCrunch, there was more than a little profane and racist commentary. On Unvarnished, these sort of things would be grounds for review removal, and moreover, a warning to reviewers to adhere to Community Guidelines, or risk banning.
But, unsurprisingly, for the other reasons discussed above and below, the number of reviews on Unvarnished that have been removed for community guidelines violation have been exceedingly small, as the community on Unvarnished is committed to professional, productive reviewing, and has the proper incentives to maintain that.
Community Norms and Standards – A pervasive concept of what is and isn’t OK

TC – Raucous vigorous debate: Anyone who’s a reader of TechCrunch knows that in part because of the audience, in part because of the subject matter, and in part because of the technology choices they have made for their commenting system, the community norms in comments there lean towards rather salty. It makes for exciting reading. But perhaps not the most nuanced, subtle analysis.
UV – Professional, Productive Reviewing: On Unvarnished, we have taken great pains to ensure that the conversation has started and stayed professional and productive, seeding the community with a core set of trusted users, and only expanding through invitation, where new users are invited by existing users “vouching for them” in the form of a review request.
The result has been a community with high reviewing standards that reflect the Community Guidelines, educates new users as to the community norms, and provides a benchmark standard to which reviews are compared by community members as they go about voting on reviews.
Community Moderation – Rating the raters

TC – No moderation tools: TechCrunch’s use of WordPress’ commenting system, which has no tools for community moderation (upvoting / downvoting / abuse reporting ) makes it such that the community cannot police itself of bad actors, and readers cannot distinguish which comments others in the community find more valuable or more problematic.
UV – Community Moderation and Abuse Reporting: On Unvarnished, the fact that the community can vote on reviews and report as abusive problematic ones is core to helping the community police itself. What’s more, this ties in with persistent identity, in that the community votes on reviews, and abuse reports, accrue to the reviewer reputation of the moderated reviewers.
The result is that high quality and low quality reviews are quickly identifiable to readers, who can judge them better, accordingly.

Content Blacklisting – No F-Bombs, thank you.

TC – Anything goes, until it’s moderated: On TechCrunch comments, commenters can use any language they please. F-bombs, epithets, and other unprofessional language included.
UV – Clear content blacklist with warnings: Unvarnished has a clearly defined content blacklist, which acts as a backstop to our guidelines. If someone tries to use a word that we’ve deemend unfit for our community, they get a polite error asking them to choose other phrasing.
Reputation Owner Enablement: Where was Madhu’s voice?

TC – Some representation, not formalized: The person being discussed doesn’t really have a place to participate, formally, in this conversation. And while it appears that Madhu may have reached out to people he thought had a better sense of his professional capacity to comment in his defense, those comments were sadly indistinguishable as valid for the reasons explored above.
Morever, Madhu simply did not have a place to respond to commentary, and exercise his own voice in the conversation. This lack of counter speech is deeply unfair.
UV – Profile owner empowerment: On Unvarnished, we ensure that the person being reviewed has a voice in the conversation—and a loud one at that—just not the only voice.
Professionals who claim a profile may respond to each and every review, to thank, clarify, or even refute. Moreover, profile owners can reach out to their networks, and ask people they feel have an accurate knowledge of their professional performance to come review them.

A missed opportunity
At the end of the day, this whole episode throws into sharp relief why this is an important area for investment, why we think Unvarnished is engaged in an important project.
Without trying to assess what indeed is the true story with Madhu, it’s clear that any reader trying to draw conclusions based on the comments would have a very hard time doing this with any certainty.
And this is unfortunate because if, indeed, these claims were accurate, if Skype’s management team were able to access credible versions of this information, from real people, that had been validated by a professional community, ahead of time, they could have saved themselves the massive cost of hiring someone only to subsequently let them go (if that is indeed what has happened).
Morever, it’s not fair to Madhu to not understand what, indeed, coworkers thought of him long ago, so he could act to correct those behaviors before it came to this juncture.
And lastly, it’s not fair to other, potentially better fits who could have gotten the role, but missed out for lack of better information to inform the decision-making process.
And if these claims are not accurate, then it’s even worse, where bad information, surfaced through a tool not designed for this purpose, was mistakenly used to rush to judgment. If Madhu was unfairly smeared through bad actors’ abuse of a tool, again, not designed for this case, that would truly be a shame.
Either way, a lot of opportunity was missed here.
This is important stuff. It takes real thinking and engineering.
Professional reputation is an important topic, with high stakes. This isn’t a question of going to a restaurant and having a bad experience because of gamed Yelp reviews. Or getting a camera from Amazon that you end up not liking. Or a hotel from TripAdvisor that you didn’t meet expectations. The topic deserves similarly weighty investment and engineering…which is exactly what we see as our mission.