Trustpilot takes down online agent’s reviews
This story on PIE prompts the question: just whose reviews can you – and your customers – trust?
We are not going to pick on any particular review site here, but we will point out some general shortcomings across the reviews world:
- Algorithm based moderation: We tried that – effectively it involves writing a computer programme to spot fake or fraudulent reviews – back in 2008, but it did not work then, for the same reasons it does not work now: human nature combined with the wonderful English language. Human nature means that if there is a flaw in a system someone will try to exploit it: the business, disgruntled clients, ‘helpful’ friends of the business or its employees, ex-members of staff. How do you programme a computer to identify all of these? You cannot. And we don’t. Our solution: human moderators and a two-strikes rule. Our experienced moderators read every review. And their experience means that they can recognise a fake positive on sight, at least nine times out of ten (all negatives, of course, are automatically subject to scrutiny through Resolution™ – pre-publication). This means that a HelpHound client’s chances of being able to post multiple fake positives, however sophisticated their attempt(s) are so slim as to be negligible. This is backed up by our ‘two strikes’ rule: any client posting a fake review gets a final warning, another fake and they are no longer a HelpHound client.
- Flagging: having an internal flagging system is not enough. Our solution: any reader of any review is able to flag that review, immediately and on-screen. If they do so the review becomes subject to intensive scrutiny, which may include suspension of the review in question until we can be assured of its veracity.
- Publishing negatives before contacting the business: this is tantamount to an invitation to game any review site. It is not sufficient to say that the business has a right-of-reply, the damage is already done. It is vital, for business and consumer, that any criticism of the business is valid, fair and balanced. Our solution: Again: Resolution™: every negative is first served to the business. All reviewers retain the right to publish, but the business is allowed to respond to correct errors of fact.
- Share price = volume of reviews: This is the cynic in us coming to the fore: if the review site is a quoted entity (or has ambitions in that direction) the stockmarket has an unfortunate habit of imputing value in relation to the quantity of reviews the site hosts. There is a therefore a massive incentive for the review business to host the maximum number of reviews. Our solution: have a business model where the business’s value is not related to the number of reviews it hosts.
If you are going to engage with reviews you must make sure that whatever mechanism you employ has credibility: we are all familiar with the headlines about sites like TripAdvisor (remember when the ASA stopped them using the word ‘trusted’?) so we need to be able to answer the basic and central question from consumers – our and your customers – “Why should I trust your reviews?”
From the above you will see that HelpHound takes this aspect of our service – credibility – very seriously indeed. Whilst there is no system on the planet that can – or ever will – provide a guarantee that every review written is cast-iron certain to be 100% genuine (is the writer related to any member of staff at the business in question, and, if so, are they mother, father, brother, sister or second cousin three times removed?!) we have done, and continue to do, everything in our power to ensure that a review written through Dialogue is the genuine article, to be relied on when booking a hotel room, choosing a financial advisor or estate agent, in fact using any business or service whatsoever.