Have you ever typed a company name into Google and been instantly greeted with words like ‘scam’ or ‘complaints’ by Google Autocomplete? That can’t be good for business, and it is a situation that more and more organizations are finding themselves in.
Being part of a firm that is heavily involved in Online Reputation Management (ORM), we take on a wide range of tasks. There is a bit of everything, from helping individuals clean up search results for their names to partnering with corporations to identify and fix reputation problems.
One of the most common issues we get now, from organizations both large and small, is to take on the task of developing and implementing strategies that will influence Google Autocomplete.
These clients look to us to help them identify the source of the problem (e.g. do they really have some bad business practices or is it an upset former employee or competitor?) and consult on how to address and fix the issues. We then develop a strategy to influence Autocomplete to highlight the positive aspects or activities associated with the brand and push negative values out.
A Data-Driven Approach
Data drives our efforts to help our clients, so a few months ago, we set out to take a deep dive into Google Autocomplete. There is some other research online about how Autocomplete may work, but we figured it might be best if we started fresh with our own data.
We began by building a dataset made up of what Autocomplete suggests for a large number of companies, and then performed analysis on this data to identify if it could help us with many of our reputation management projects.
Consistently hearing from businesses with similar Autocomplete problems helped us hypothesize that by looking at the Autocomplete values for hundreds of companies, we could identify a list of values that Google favors when providing Autocomplete suggestions to searchers.
Our intention would then be to use these ‘favored values’ in our efforts to influence Google Autocomplete for other brands and larger companies.
While we won’t go into exactly what we do to promote a specific word in this article, it is worth noting that based on our testing, we believe the Autocomplete algorithm is comprised of 3 main influencers:
- Search volume and searcher location – the amount of searches performed for a keyword along with the location of the searchers
- Mentions of the keyword on the Web, crawlable by Google’s spider
- Social Media mentions of the keyword on sites like Twitter, Facebook, and Google+
If you can obtain these 3 items in large quantities for your desired keywords, you may be able to influence Google Autocomplete.
If you are interested in learning more about white drives Google Autocomplete, this Google Instant/Autocomplete article provides a strong background on the subject, and Google also pulls back the curtain a tiny bit on their Autocomplete support page.
Methodology
The first thing we needed to do was identify our starting point: what companies do we look at to determine if there is an Autocomplete bias?
We decided that we would use the Fortune 500 to get a list representative of very large companies. We then also chose the INC. 5000 to complete the dataset, as this list contains companies ranging in size from very small to medium/large.
The companies selected from the Inc. 5000 were limited to the 500 at the top and 500 at the bottom of the list. The top would represent mid to large sized businesses, and the bottom would be the small businesses.
We analyzed them all separately and broke out the data so that we had findings that pertained to small business, medium-sized businesses, and large businesses. We noticed that there was some overlap between the top 500 on the Inc. 5000 and the Fortune 500, so for the top of the list, we skipped the first 100 on the Inc. 5000 and grabbed the 500 that followed.
Next, we used an undocumented Autocomplete API from Google that displays the top 10 values for any given word or phrase.
This is the format that the API uses (you can access it right in your browser):
http://google.com/complete/search?q=YOUR+PHRASE+HERE&output=toolbar
We used some Google Docs magic to scrape the Autocomplete values for the 1500 different companies and then dumped the data into Excel for some more sophisticated data crunching.
What This Raw Data Looks Like
Want a visual representation of what the Autocomplete value results look like? This Wordcloud (thanks Wordle.net) is made up of the 13,500 Autocomplete values we pulled (9 values for each of the 1,500 companies).
Once we had the data in Excel, we were able to analyze it and find data points such as the most frequently appearing values by occurrence, differences between top values in large and small businesses, negative keyword occurrences, and more. We’ve dug through the data to find the most interesting and actionable information for you.
Autocomplete Analysis For The Fortune 500
The Fortune 500 is made up of the largest companies in the US by gross revenue. You can view the 2011 list of companies that we used here.
We first analyzed the Autocomplete values for these companies, and the following are the top 10 Autocomplete values for them based on occurrences (occurrences follow each of the words in parenthesis).
- Careers (227)
- Jobs (153)
- Wiki (145)
- Investor Relations (140)
- Stock (108)
- Locations (87)
- News (72)
- Foundation (52)
- Coupons (42)
- Headquarters (40)
It is interesting to note that ‘Careers’ and ‘Jobs’ occupy the number one and two spots in frequency.
Our takeaways on the prominence of jobs/careers in Autocomplete are:
- Your company should make an internal decision on what you’ll call your employment area. Will you call it ‘jobs’ or ‘careers’? Whatever it is, be consistent with it across your organization.
- It is clear that people search for company name + jobs/careers a lot. Make sure you have a well-built out area on your site catering to job seekers. Also consider syndicating your job listings to sites like indeed.com, careerbuilder.com, snagajob.com, etc so that you build a strong correlation in Google’s algorithm between your company and the concept of employment.
- Encourage candidates interested in work to search for “Your Company name + jobs/careers”. So, instead of sending someone a link to your job listings page, tell them to just Google “Brand Name Careers” to see all of your listings. The searches this generates signals to the Autocomplete algorithm that people are interested in that search phrase.
Here are a few more quick-hitter thoughts on the top 10 values and how you could use this data for your own brand:
- Wiki – Consider starting a Wiki that is associated with your brand name. If you are a big business, you can probably create a wiki all about your company. If you are smaller, it may make more sense to build a wiki about your niche or industry.
- Investor Relations – You don’t necessarily need to be a public company to have ‘Investor Relations’. Consider creating a section for this on your site, with resources such as press releases, letters from the CEO, a news feed, contact information, analysis reports, links to important blog posts, etc.
- Locations – If you have multiple locations, build out a dedicated page for each location on your site.
- News – It may be old school, but have a ‘news’ section on your site if you can keep it up-to-date.
- Foundation – Consider being a do-gooder and start a foundation. Many organizations already do things to give back to the community and society, so why not consider formalizing that a little more through a foundation.
- Coupons – People are always searching for these. If you are in a business where these make sense, take control of it and develop a strategy to use it to your advantage.
These top 10 words appear to be very popular in Google Autocomplete for the Fortune 500. In fact:
Other Interesting Notes From The Fortune 500
- McDonald’s shows ‘Coffee Lawsuit’ in the 9th position, nearly 18 years after the landmark coffee trial took place. It is clear that negative values just don’t go away on their own, and that you must actively work to keep them from displaying for your brand.
- There are quite a few companies that have negative Autocomplete values lurking. Google displays just 4 Autocomplete values for most users, so values in positions 5+ don’t display. Pfizer, Lockheed Martin, News Corp, New York Life, Qwest, Bristol Myers Squibb, and a bunch of other well known brands have negative values such as ‘lawsuit’ or ‘layoff’ lurking in those deeper positions, but they could rise up at any time. These companies should be actively working to push these negative values out of their top 10, because if they don’t, a negative value could move up into their top 4 and cause some serious branding issues.
Autocomplete Analysis For Small To Mid-Sized Companies
It also made sense for us to analyze companies that are quite a bit smaller than those on the Fortune 500. This would provide results that would be more meaningful for small businesses. We decided to use the Inc. 5000, which is comprised of the fastest growing companies in the US.
By grabbing the companies at the end of this list (500 of them), we could pull a good representation of what the Autocomplete landscape is like for small businesses, and what other small businesses could do based on the data gathered.
The top 10 values we found, based on occurrences, are:
- Inc (99)
- LLC (66)
- Jobs (57)
- Reviews (54)
- Review (49)
- Facebook (28)
- Coupon (26)
- Blog (23)
- Address (23)
- Careers (23)
It is important to note that a blank value was, by far, the most frequently occurring value, but we didn’t show it in this list above. This means that a value was not returned in one or several of the positions for a company.
Key Takeaway For Smaller Enterprises
If you are a small business, use the analysis in this document to define and influence Google on what your Autocomplete values should be. If you wait for Google to fill-in your Autocomplete values, you could end up with brand name + ‘scam’, ‘complaints’, or worse.
This slide deck has some proactive tips you can take to protect your brand’s online reputation.
These are the other highlights from analyzing these smaller companies:
- Jobs/careers again proves to be ultra-important, so follow the advice outlined in the Fortune 500 section for this topic.
- Taking control over your reviews is important, as it seems almost inevitable that review or reviews will be associated with a company in Autocomplete. Make sure you are stockpiling positive reviews on Google Places, Yelp, industry review sites, and your own website, especially before you think you’ll need them. There is nothing like a negative review to throw everyone in your company into a frenzy.
- Facebook has a strong showing, and so you know it will only become more and more important that you have an active presence on the social network. I’d imagine that when we run this exercise again in 12 months, with Google pushing Google+ so strongly, we’ll see Google+ somehow “work its way in” as well.
- While not in the top 10 list, we did find many occurrences of ‘lawsuit’, ‘scam’, ‘layoffs’, and other unsavory recommendations. We won’t out any small companies here, but we do advise people to look at their Autocomplete values often. They change about every 6 weeks, and you don’t want to be caught off-guard with a negative value that can cripple your branding. We provided the API link format above, or if you want an easy input box you can use the one we’ve built here.
Autocomplete For 500 Mid to Large Companies On The Inc. 5000
Following an analysis of the 500 at the bottom of the Inc. 5000 list, we moved up to the top of the list and grabbed 500 from there (after skipping the first 100 due to overlap with the Fortune 500).
These represent mid to large sized companies who have shown tremendous growth over the past several years. For these companies, the top values occurring in Autocomplete are:
- Jobs (119)
- Careers (110)
- Reviews (76)
- Salary (62)
- Wiki (56)
- Inc (46)
- IPO (41)
- Address (34)
- LLC (31)
- Revenue (30)
What is interesting about this is that it shows elements of what we see in both the Fortune 500 list as well as the list from the smallest companies on the Inc 5000. This is evidence that a brand’s Autocomplete values shift as the company grows.
Here are some notes on the new values we see:
- Salary – be aware that, for larger organizations, people will be searching out salaries. You’ll want to take ownership of the results, and make sure that you are representing your brand positively in those results. There are many sites like Glassdoor.com and SalaryList.com where most medium and large-sized companies are listed, and you’ll want to review how your company is represented on those sites.
- IPO – I can only imagine that this shows up if there are talks that the company might be going public or has already gone public. This is another place where it pays to check your results. If there is negative content out there for it, develop a strategy to turn the discussion in a more positive direction. Negative Branding in any form could impact IPO or future stock prices.
- Revenue – People will be searching this out. Have a section on your site dedicated to it, as it will rank at the top, and you’ll be able to control the message.
- We also noted a significantly large portion of the Autocomplete values were city specific, typically the brand name + a large city name. This demonstrates that if you are present in multiple cities, they can be used to fill out your Autocomplete values.
Combining All Of The Data
While it was useful to segment the companies by their source (Fortune 500, Top of the Inc. 5000, Bottom of Inc. 5000) to understand the differences in Autocomplete values across different business sizes, we also stacked all of the data and analyzed it as a whole.
You saw the tag cloud at the beginning of this article that was the result of all of the values combined, here is the overall Top 10 list – at this point, there shouldn’t be any big surprises:
- Careers (356)
- Jobs (329)
- Wiki (207)
- Inc (162)
- Investor Relations (150)
- Reviews (140)
- Stock 130)
- Locations (109)
- LLC (98)
- News (95)
Categorizing The Values
After uncovering the Top 10 values, we then assigned categories to any keyword that appeared at least 5 times in the stacked data. By doing this, we could see which category of keywords appears most frequently overall.
The following are the top 10 categories:
- Locations
- Employment
- Investment Information
- Company Structure (LLC, Inc, Co., etc)
- Social Network/New Media
- Complaints or Reviews
- General Information
- Coupons
- News
- Negative
What is interesting here is that while we have seen career and jobs dominate the top of the lists, when categories are assigned, ‘Locations’ is actually far more common. This is because ‘locations’ is made up of so many different values (the word location, locations, plus every city and state reference).
Business owners should also take note that ‘Complaints or Reviews’ is the 6th most frequently occurring category, and the negative value category (scam, lawsuit, etc) is the 10th most frequently occurring. Both of these areas are hotbeds for negative branding activity, so it pays to be proactive in maintaining positivity in these areas.
What Do We Do With All Of This?
Overall, what we have learned is that there isn’t nearly as much variability in Google Autocomplete as we first thought coming into this. There is clearly a correlation between keywords used for businesses and the values displayed in Google Autocomplete.
For a company proactively trying to take control of their Autocomplete values, the data we compiled into the top 10 lists provide a good start for the values to target.
Once you finalize a list of your target positive values, it just takes some ingenuity and clever thinking to get people searching those terms, mentioning them in social media updates, and including them in content published to the Web. A few months of that, at a decent volume, is typically enough to sway Autocomplete in the direction you want it to go.
Limitations
The data isn’t perfect, but we think it is strong enough to make decisions from when used in aggregate.
Some limitations include a location bias in the Autocomplete values (based on where the Google Docs IP address geolocates) and the fact that values for some companies are very specific (e.g Boeing where most values are plane models like 747, 757, etc) and one-offs like that don’t show up in reporting although it is important to note that some companies have this.
Also, because we input the company name directly into the Autocomplete API, we only received nine usable Autocomplete values because the first result was always the company name.
Next time around, we’d insert a space after the company name, as this will generally remove the company name as the first recommended value and would give us 10 values to fully analyze.
Is There Autocomplete Analysis That You’d Like to See?
We think that there is a lot more research to be done with Google Autocomplete, and we’re going to keep exploring. Robert Darlington, Dan Hinckley, and myself spent quite a bit of time figuring out exactly how to pull this experiment off, and now that we have the tools in place, we’ll be pulling even more data and performing even deeper analysis.
Is there something you’d like explored with Google Autocomplete? Let us know in the comments!
Opinions expressed in the article are those of the guest author and not necessarily Search Engine. Land.
Article Source :- http://searchengineland.com