New Google Patent on Reputation Mangement
Do reviews of businesses and products at Google influence how well those might show up in place searches or product searches? They may play a role, and a bigger question might be how much weight might Google give to each review that it sees. An answer, in part, to that may depend upon rater and reviewer reputations scores associated with the people leaving reviews.
People do go online to search for reviews and ratings for businesses and products, and the search engines are trying to provide that information when and where they think it might be useful.
Starred ratings are also showing in Google’s Web search with Rich Snippets, and the presence of ratings may influence whether or not someone clicks through a snippet from Google’s search results.
A recent change to how Google shows search results in Web search may mean that if Google thinks you are performing a search where local search results are appropriate, then Google may show those local results as if they were organic search listings. Google refers to this change as Place Search, and it can have an impact upon the number of visitors a site may receive and possibly increase the number of contacts for a business listed in those results.
A patent from Google granted today, Systems and methods for reputation management (US Patent 7,827,052), takes a closer look at the people who provide reviews and ratings for businesses and products, and describes a way of creating reputation management scores for them modeled after Google’s PageRank algorithm.
The reputation management patent was originally filed in 2005. It’s possible that Google is using an updated version of the approach described in the patent, or adopted an entirely different approach. The patent, as it’s written, seems to apply primarily to reviews and ratings that people might make to Google, but could potentially be expanded to include reviews and ratings from other sites as well.
Google has looked more deeply at just ratings, and last year, I wrote about a Google patent filing that describes How Google May Rate Raters, which seems to share a few ideas with this newly granted patent but looks at ratings provided on sites other than just Google.
Google has also looked more closely at the value of reviews, and I’ve written many posts on that topic as well, some of which echo the “reputation score” approach described in this patent. I mentioned this particular patent back in the days when it was still pending as an application in a post describing five related patent applications published simultaneously, Google Reviews: Reputation + Quality + Snippets + Clustering.
Some other earlier posts on reviews and ratings from SEO by the Sea include
- Google’s New Review Search Option and Sentiment Analysis
- Opinion Summaries in Google Maps Reviews
- Google Approach to Making Online Ratings Easier…
- Innovating Product Reviews at Google
The abstract from today’s granted patent tells us about:
A reputation management system, method, and computer-readable storage medium assigns reputation scores to various types of entities including, but not limited to people, products, advertisers and merchants.
A reviewer reputations function is based on a directed graph including the reviewers and the reviews. The nodes in the graph represent the reviewers and the reviews and the links in the graph represent the ratings.
The reputation function is iteratively solved until a convergence condition is met. Before convergence, when a stability condition is met, the reputation function is modified to remove portions of the function corresponding to nodes with negative reputations.
Upon convergence, reputation values for at least the reviewers and reviews corresponding to nodes that have not been removed from the reputation function are generated.
Some interesting tidbits from the reputation management patent
How might Google handle reputation management when it comes to things such as ratings and reviews? I’ve pulled out some interesting highlights from the patent.
- Some Reviewers or Raters can have negative reputations, and their reviews and ratings may not count in the final scores for the things, organizations, people, and other raters or reviewers being scored.
- When reviewers provide only negative reviews or ratings, their contributions may not count.
- In addition to creating a reputation score for reviewers and raters, this system includes reputation scores for reviews themselves.
- This system attempts to anticipate the possibility of people attempting to manipulate it. For instance, if someone’s reviews are highly rated, their reputation score may increase. But, the reputation score for that reviewer depends upon the reputation scores for the people rating his or her review. If their reputation scores aren’t very good, then their positive ratings won’t have much of an impact. Likewise, someone receiving poor ratings for their reviews wouldn’t have their reputation score being affected tremendously if those raters had low reputations.
The reputation management patent does go into considerable detail about how they may calculate reputation scores for reviewers, raters, and rateables – the people, places, and things being rated.
If you have a business that may show up in local search results or Google’s Place Search results, you may want to spend some time with Google’s patent filing on reputation management.
Added November 3, 2010, – Reviews and ratings on Google are shown in rich snippets for businesses that may be listed in Google Maps, as well as for some product sites who applied to be able to show Rich Snippets. A post at the Official Google Webmaster Blog posted yesterday, and updated today, Rich Snippets for Shopping Sites, describes how Google has expanded Rich Snippets for eCommerce sites.
It’s worth a read if you offer ratings and reviews on your eCommerce site.
51 thoughts on “How Reputation Management May Work at Google”
I think Google should use some sort of “review of raters” or else we run the risk of competitors negatively ranking anyone else for their own gain. Its current system is certainly less than perfect and leaves a lot to be desired.
First and foremost, they need to address the review farm industry. You know, the 67, 5-Star reviews a local business paid someone to post on their behalf. It makes a mockery out of the entire concept of the customer review platform, Google, Yahoo, Insider Pages or otherwise. The there are the phony competitor reviews along side the fake reviews posted by the owner. Looks like this patent falls a little short of efficacy.
I actually seek reviews when I’m actually planning to buy a particular product, through reviews I can gauge if that particular product is good or not. At least now I can rest assured that Google is doing the best they can to make sure that reviews stays valid and trustworthy enough.
Great. As far as I know the methods that Google uses are recognize spammers are REALLY bad, especially in local businesses. There are tons and tons of complains about the rejected and suspended businesses in their forums. So now they will try to create a magic way to rate reviewers without giving great weapon in the hands of the spammers? No freakin way.
This is probably a lot more valuable than some people might think. Going off now to read the patent to see what else can be derived from it.
Thanks and best rgds
I had a gut feeling that there was more to the reviews and push for local recently – seems that I was sorta right – thanks for this Bill, not being a patent junkie, I know I can rely you on pulling the right info from G patents… Saved me looking 😛
I think Google is probably using what’s described in the patent on some level right now. I agree that it probably isn’t working as well as they might like.
I often check reviews before I buy something too. Sometimes I do it in the store using my phone.
Thank you Richard.
Good to see you.
This patent really doesn’t provide a way to filter out reviews that might be from competitors or people manipulating the system. Instead, it trys to find a way to give a reputation score to reviewers and raters. Those reviews and ratings may influence how well a business or product might show up in a search, but they won’t necessarily cause untrustworthy reviews from being listed.
Google does let you rate reviews, with the following appearing after reviews left on Google:
Was this review helpful? Yes – No – Flag as inappropriate
Rather than removing the reviews, they let people who see them vote upon them.
There are problems with Google Maps, and that really shouldn’t be much of a surprise – it’s a massive undertaking.
As I noted in the post, the patent was originally filed in 2005, so it isn’t something that is necessarily new.
As i see it, this could work, but still it won’t point out how true the reviewer is. Unfortunately, on Internet it’s easy to pretend that you are someone else.
Thank you. Google does seem to place a lot of importance on showing reviews and ratings, whether for local or for products.
I’m wondering how much of an impact having starred reviews showing up in snippets from more ecommerce providers will have on traffic.
Ratings and Reviews are an increasingly important part of eCommerce. I know that personally, I check hundreds of reviews before making a major (and sometimes minor) purchase. It would be good to know that there is some sort of system in place that ensures that the people giving the reviews maintain some sort of credibility. Otherwise, it drastically cuts down on the help provided by the review itself….
I look forward to seeing how this evolves.
I think it may have a result towards the local listings Google places which are now being merged with organic results. In general, it’s the “citations” that affect the local listings and that does include reviews, among other factors.
Regardless of whether someone might make reviews under a pseudonym or not, it’s possible that if that person doesn’t leave very many reviews, they may not attain the kind of reputation that might make much of a difference when the value or weight of his or her reviews are considered.
I don’t mind looking at reviews before making decisions when it comes to purchasing products or services. I spent a lot of time over the past couple of days looking at a lot of reviews on the Amazon.com website. One of the things that I really do like is the “did you find this review helpful” option for those reviews. I also found myself looking at a lot of choices that had reviews over ones that didn’t.
I think I actually liked the seven box (or older ten box) better than the display of local listings as if they were organic search results.
Citations, or mentions, are an important part of how a local search result might rank, but not the only consideration. I think reviews can be more important for some queries over others – some business types just don’t seem to generate too many reviews.
I think distance still plays an important role as well, for some queries.
@ Jim Ryan & Plamen
Agree 100% from the working side.
But, as Bill mentions “.. describes a way of creating a reputation score for those reviewers” : NOT THE REVIEWED.
So maybe spending more time building out whatever is worthy to enhance ones reviewer reputation might be more productive, faster, more beneficial for the client. I already think of Bill as an A-List Blogger! So this might forecast the A-List Google Reviewer.
So it’s like enormous amounts of pathetic links vs. several, highly selected, droolish types with exact-match content possibly.
At least I will pretend this might be the case.
As usual providing as much personal/business information to Google from within the account as possible is a nice way to begin the buttox puckering. A simultaneous dose of imagining your an editor within DMOZ by spending a fair amount of time on sites which you have no detectable historical link relationships might also help your ascent toward attaining A-List Google Reviewer status.
I already think of Bill as an A-List Blogger!
That IS an OLD patent. What is up with that?
How often do Google work on their algorithm? It’s always difficult to stay on top of how Google works because from what I understand they swap around their algorithm all the time. If you’ve put all your eggs in one basket and that basket becomes null – then you’re screwed.
It is an old patent, and it’s possible that Google has been using bits and pieces of it for years. Over at Understanding Google Maps & Local Search, there’s been at least one recent post that points to Google stepping up their efforts to police reviews that’s worth a look:
Google Places: Google Confirms New Review Removal Practices
Google’s work on its ranking algorithms is an ongoing process. A number of search engineers from Google have stated over the past year that they often make at least one change to how the search engine ranks pages daily. Some of those are very minor changes, while others are very significant.
The ultimate goal is the same regardless of how the algorithm might change – help searchers find relevant and important sites in response to their queries.
I could only hope that Google would take the reputation of the reviewer into account. I am finding quite a bit of dirty reviews lately from competitors and in some cases there is no recourse. At least this can tamper that a bit.
It appears that Google is attempting to take the reputation of reviewers into account with the approach described in the patent filing.
Google Hotpot should (and lets hope will) elimiate some of the comment spam that we see on Google places listings.
The new profiles for Google Hotpot might work well with a reputation system like the one described in this patent, where different reputation scores for reviewers can influence the potential impact of those reviews.
It seems it could work, but still will be doubts. I think if Reviewer A has a history of writing both positive and negative reviews that other people have marked as helpful, I imagine their reviews are going to count more than those of Reviewer X, who comes out of nowhere to announce that some company is the worst thing in history.
The past history of reviews gives the search engine something to base a reputation score upon, so I think you’re right with that.
If Reviewer X continues to write a number of other reviews, including some that are positive and some that are negative, and other people mark those as helpful, then it’s possible that their initial review will carry more weight.
Thanks for pointing me to this article. Interesting read and since your original post, Local Business Center / Google Places has definiately undergone a lot of (experimental??) changes. Some of my clients seem to get most of their traffic through their Google places listing. It’s a little hard to tell sometimes, but since I am only a small business I am able to look at their search traffic quite regularly and search phrases are often bringing up an “organic” Google Places listing rather than the actual site listing.
Sorry – was mixing the date formats up!!!! Had a brain freeze and suddenly though I was look at a reply from 11th March rather than 3rd November. Sorry – that’s our Brit date formats for ya!
I’ve worked on some sites where local search become one of the most important ways of drawing new visitors to a page. For some types of businesses, it’s a more powerful attractor than even their own websites.
Great post . . .
Any ideas on how the “flag as inappropriate” tag will really work?
I am not that familiar with the recommendation piece on Google Local yet – BUT just thinking if I have a small business and I have some “less than honest” competitors leaving false or misleading recommendations/comments can I just remove them? Seems just checking that the recommendation “wasn’t helpful” still leaves the same bad taste in a potential customer’s mind.
This highlights the fact that I have recently seen a new crop of “flagging services” begin to pop up in the world of Craigslist marketing – where the same folks who were getting paid to do manual postings are now branching into not only hammering out multiple postings for clients (obviously against CL rules), but also now flagging their competitor’s posts as well in an attempt to get the competitor’s posts removed or at least causing enough confusion that the competitors quit using Craigslist as much or maybe even all together. And believe it or not Craiglist has become a reasonably profitable piece of many Realtors and Car Dealers marketing activity over the last few years.
Seems like Google has some WAY more powerful options that are far less open to manipulation to rank local sites than random “recommendations” – am I missing something? With say even Ebay’s ranking system – a competitor would have to actually purchase an item – then falsely bash their competitor’s ranking after that purchase for “poor service” – a comparatively expensive and time consuming strategy for dinging your competitor’s reputation. Now it could be as simple as posting a bunch of bad reviews of your local dentist and you could potentially disrupt their local ranking??? Ouch . . .
Thank you. As I mentioned in this comment, it’s possible that Google might allow others who read a review to rate those reviews. The review and rating history of people rating reviews might influence whether or not those ratings are even displayed, and how much weight they may carry. A “flag as inappropriate” might force Google to have a human being take a look at a review listed.
If someone creates a new Google Account, votes down 20 or 30 reviews in one market, while leaving a positive review for another, it sends a signal to Google that something foul might be in play. That’s the purpose behind this patent.
There’s a lot of speculation that reviews for businesses carry some amount of weight in the rankings of pages in local search, but chances are that reviews are possibly one of the weaker signals in Google’s local search behind other things such as how relevant the kind of business is for the query used, how prominent a business is in the location searched, and how distant the business might be from a center point in that location.
This is really interesting. Correct me if I’m wrong, but this means that the prominent factors in the review that is displayed on Google search are: (1) the star rating of the business being reviewed, (2) the reputation score of the reviewer who reviewed the business, (3)the reputation score for the specific review, right? (I’m sure there are other factors involved, but I’m curious about how exactly Google is choosing the review that will be displayed)
If this is true, then businesses are, now more then ever, going to need to manage their 5-star rankings and monitor their reviews in a comprehensive manner. This might also change the nature of fake reviews/fake 5-star raters.
The focus of this patent isn’t so much on how Google might rank reviews themselves, and decide which ones to display, but it does show one aspect of that process – how much trust it might place in reviews and ratings themselves from people who submit them.
It’s possible that when Google decides which reviews to show, it may also consider sentiments expressed in those reviews. See: Opinion Summaries in Google Maps Reviews.
There’ve been lots of fuss lately as bad reviews by competitors hurt company image. At this point I dint think reviews have any power except maybe to engage a prospect.
I have seen a number of instances where client reviews (credible reviews from review sites – not the new breed of paid review spam) have made positive impacts in local search rankings. I push (hard) getting real live reviews from clients in the office with almost all local clients.
I have not yet seen any obvious negative ramifications from spammy reviews though . . . it does seem that Google somehow places higher value on reviews from more credible review sites, but in many ways like SEO it is not exactly black and white if say 5 reviews from some shady paid review site trumps one direct review on Google from a busy Google+ signed in user? It does seem like it would be rather simple for Google to just disregard reviews in their algorithm from sites with less than a certain page rank or that don’t require some higher level of hurdle for users to leave reviews. Now drilling down to the specific USER review level (the basis for this patent) does seem pretty impressive.
I have always found in interesting how quickly small niche markets build up around new and different online marketing strategies. Buy facebook friends or likes or Twitter Followers . . . Cragislist Ad Posting/Flagging . . . Paid Review Sites . . .
I’ve seen at least one instance of the positive impact that a review can have in helping a business rank in local search results.
Regardless of that though, people seem to want to read reviews, and Google is trying to make it as easy as possible, even including reviews where appropriate in the new Google feature, Google Related.
I’ve seen some instances where reviews have had positive impacts in local search rankings as well.
I suspect that for spammy reviews, the weight or value of those and other reviews from the same authors may not be initally worth much if the reviewer doesn’t have much of a past history of leaving reviews, and might carry even less weight if those authors continue to leave reviews that could be interpreted as spammy.
I think the genie is really out of the bottle with local search. Already there is an entire SEO niche industry based on it. It will only continue to grow. Personally, I think it’s great. I know it’s helped me in my own personal queries by delivering exactly what I was looking for locally.
I’ve been doing local search optimization for a number of years now, and it can really make a difference for businesses that have locations that they either provide services to, or have an office or shop that people can visit in person.
I’ve also found it pretty useful as a searcher as well.
Prior to this read I hadn’t put much thought into how reviews affected my seo rankings. Thanks and I will put it into practice on my local seo results.
While the patent I wrote about really didn’t go in the direction of telling us how reviews might influence search rankings, it did give us some ideas about how Google might be reviewing and placing weight upon rating activities from individuals. Given that a lot of people do go online specifically to look for ratings and reviews, I think that’s pretty useful to know itself.
Comments are closed.