Sign up for weekly recaps of the ever-changing search marketing landscape.
Our SERP Sentiment Score For A Person’s/Brand’s Online Reputation
Having trouble showing progress when it comes to online reputation management (ORM)? Columnist Brian Patterson explains his metric for tracking ORM campaign progress over time.
It’s something all companies fear: You Google your brand name, and right on the first page are negative Google reviews, a low Yelp score, damaging articles or — possibly worst of all — a Ripoff Report post.
The thought of this setback is enough to make any good marketer break out in a cold sweat. But for many of us, this is just another SEO challenge we face, albeit under a slightly different name: Online Reputation Management.
Imagine you are the marketing genius behind the iconic Snuggie. You Google your brand, see the following results and are rightly concerned:
In these results, you can see a few clearly negative search results from big news sites (boxed in red). There are also two blog posts that you might consider neutral (boxed in blue). These neutrals aren’t the worst things in the world, but they aren’t glowingly positive, either.
Just two negative articles appearing in the search results for a product can cost you up to 44 percent of potential customers, so this is definitely an issue to address, as it could be significantly impacting online sales.
Is It Off Page One?
As with every project, we want to start by capturing a baseline starting point and determining how we will report improvement over time.
Tracking progress for an online reputation management (ORM) project like this can be complex. Instead of tracking a single domain for a large set of keywords (as with SEO), we are tracking multiple URLs for a smaller set of branded keywords. So while SEO metrics are well established, ORM metrics are almost always custom-defined for the client and project.
The initial focus of a reputation management project with a client is almost always, “Is it off page one?”
While that is a valid question, completely cleaning the search engine results pages (SERPs) for a brand name can take months. And in that time, if you are only focused on a binary yes/no result, you miss out on how improvements or setbacks are impacting your brand.
This binary question misses some of the nuances of the search results (namely, their ranking positions). As we know, ranking position impacts user visibility and click-through rate.
Because of this challenge, we built a SERP Sentiment score to measure branded SERPs based on the number of positive, negative and neutral URLs that appear on page one, along with what positions each of those occupy.
This score helps you see and report the progress of your reputation management campaign on a daily and weekly basis — the type of progress that isn’t noticed if you focus on the binary result of, “Is it off page one?”
The 100-Point Algorithm
We researched, discussed and tested different approaches to calculate what impact negative SERPs should have on our sentiment score. We broke through when we started with 100 points and deducted points for every negative or neutral search listing on page 1.
Each ranking position has a preset value based on the (admittedly imperfect, but useful) average CTR distribution.
For a SERP with 10 results, the value of each position looks roughly like this:
Calculating The Score
Each SERP starts with 100 points. If a piece of content in a position is negative, it loses the value of that position. If the piece of content in that position is neutral, it loses half of the value of that position. And if it is positive, no points are lost for that position.
In the example above, everything is positive except for position 8. So we would start with 100 points, subtract the value of position 8 (value of 4), to give a SERP Sentiment score of 96. Not bad, Ron Popeil.
If there was a neutral result, we’d divide the value of that position in half, and subtract it out. However, in this case, there is only one negative result in the SERPs to contend with.
Let’s calculate another:
In these results, I noted the positions where we need to subtract points. Negative content is again boxed in red, neutral content in blue. We start with the 100 points, and then adjust as follows:
- -9 points for the neutral result in position 2 (This position is worth 18 points, and having a neutral result there means half of that is subtracted).
- -8 points for the negative result in position 4.
- -2 points for the neutral result in position 8.
After these adjustments, the SERP score for this query is 81.
Scoring Non-10-Blue-Link SERPs
When we built this score, 10 blue links was the rule, and Google did not vary from it. However, now the number of listings in SERPs can vary widely, with things generally settling between seven and 11. In those cases, we modify the scoring as follows:
Using the scoring for nine SERP listings, the SERP Sentiment score for Snuggie’s brand results (from the beginning of this article) is 81.5.
Tracking & Reporting
We first establish with the client what is positive, negative and neutral. We read through each page, assess it and make a recommendation on the sentiment (but at the end of the day, it is the client’s call). We also establish which queries we are going to monitor. Generally, it is the brand name (or person’s name) and also queries such as “[brand name]+reviews” or “[brand name]+product name.”
We track weekly the search results for each query, monitoring five pages deep, so we are aware of anything moving up from deeper in the SERPs. The SERP score, however, is only based on page 1, since the vast majority of searchers never click to page 2.
For ORM projects, we report on the ranking position of each piece of content. (We have a tool that automatically tracks ranking positions, but it can be done manually.) We also report on this SERP Sentiment score, which is a simple way to roll up these positions into a single number. We provide weekly and monthly reports, which go into detail on the work completed and its results.
Since developing this sentiment score, we’ve found our clients’ C-Suites to be more engaged and interested in incremental progress, because they can clearly see the project direction through the changes of a single number. They know that as their score gets closer to 100, things are getting better.
Furthermore, with the utility that we’ve found in this score, we’ve also built out other ways to score online sentiment that factor in online reviews (both the aggregate rating and the number of reviews), knowledge panels, Google images and more. Scores and dashboards really add value and make tracking improvement easy for even those only casually involved on the project.
Google changes their search result pages constantly, which makes this sentiment score a work in progress to keep up to date. I hope that by sharing this metric, others will be able to use it in their own projects and improve upon it (and ideally, share their updates).
Customer satisfaction has the Net Promoter Score. Now, online reputation management has the SERP Sentiment score.
Some opinions expressed in this article may be those of a guest author and not necessarily Search Engine Land. Staff authors are listed here.