If you’re not filtering out affiliate spam sites from measurement results, it could be.
Affiliate spam websites and Twitter handles are built to drive traffic to sites like Amazon with an end goal of earning a percentage of the sales dollars. These sites regularly “borrow” content from directories in order to bolster their search rankings. While Google and other search engines have made progress in the battle to filter out or push down these sites in search results, many social media aggregation tools are still returning a number of affiliate sites in results.
What does this mean for marketers?
When conducting online measurement and research for consumer packaged goods, we’ve often found that affiliate sites make up 40-60% of social media results. For some clients, this means that results may appear to be thousands of posts higher than they actually are. At Loudpixel, we tend to reference the percentage of affiliate to “true” opinion posts and reactions, but it’s important to exclude these sites from any in depth analysis—this overly promotional content will skew the numbers.
And to make matters more complicated, tools such as Radian6—which work on tiered pricing structures based on volume—could be more you costing more than necessary to pull in extraneous content.
How to identify affiliate spam websites
- Check out the URL. If you see that the main URL is a string of mumbo-jumbo letters or numbers (e.g. ab3bg.com/widgetreview) or directly related to selling a particular product (e.g. buywidgets.com/widgetreview), there’s a good chance you’re looking at an affiliate website
- Look for multiple posts with the same content. Retweets are one thing, but if you see a string of the same overly positive or promotional posts from different sites with no direct attribution, you’re likely looking at affiliate spam
- Look for specific keywords: Proceed with caution when you see words like “buy cheap,” “buy now,” “low price,” “hot deals,” “online shop,” “Amazon” or other content that is overly promotional toward the brand or company—in fact, in many cases, you can filter these words out from the get-go, depending on the specific client and goals
- Look for a string of positive reviews followed by links to Amazon. Affiliate sites will often repost positive product reviews promoting the product, followed by a link to buy the product online
- Use your best judgment. Sometimes, affiliate spam is obvious (as in the example below), but the more you pay attention to the general design and features of the sites, the better you’ll be at identifying them
In order to return accurate measurements and results, take the time to identify and familiarize yourself with the characteristics of these sites. Otherwise, you’re probably wasting effort.
Posted in Culture by Jeff | No Comments
In a perfect world, everything can be accomplished by the touch of a button. Alas, this world is far too complicated to rely on buttons—it requires the human touch in order to keep things running smoothly.
This is the case with social media analysis. While computerized monitoring tools can provide solid quantitative data, they are often unable to pick up on the humanistic side of social media responses. For example, a data source may read, “I’m so glad I spent 10 bucks on this awesome movie….right…” as a positive post. However, most humans would classify this as a sarcastic comment. Monitoring software can easily misread sentiment in sarcasm, ambiguity, and may likely incorrectly categorize simple positive or negative posts.
Sentiment is just one aspect of social media analysis that can skew product perception when it is solely computerized. As we’ve reiterated time and time again, human review is necessary in order to guarantee accuracy of information and to fully utilize the capabilities of whatever software you choose to employ. These online tools are essential for securing data quickly and efficiently, but in order to get the most out of your monitoring program, you’ll also need a good pair of eyes.
Posted in Tools by Lesley | No Comments
The race is on to find the perfect social media monitoring tool. A number of big players, including SAS, IBM and Microsoft, have entered the playing field to compete with some of the original players, such as Radian6, Biz360 and Scout Labs.
But the fact is, no single tool is perfect. Whether dealing with the difficulties of Facebook privacy issues, capturing accurate sentiment, filtering out spam or trying to pull in the most relevant conversation data for the brand or organization, no tool can do it all. The perfect technology solution for social media monitoring simply does not exist.
This means that users of such tools need to be extremely critical of the conversation data they’re using. While the tools are powerful—and a great start to drawing insightful conclusions—users need to be checking for irrelevant posts, constantly updating keyword search strings, considering the best approach to tracking sentiment (human vs. automated), keeping an eye out for content duplication and watching some conversations (brand Facebook walls, for example) more closely on a manual basis.
Don’t blindly trust your conversation data. To use a simple metaphor, post data should be approached like a garden—weed daily and apply regular maintenance to avoid becoming overrun with content you don’t want.
Posted in Tools by Allie | No Comments
The concept of listening to social media conversations has caught on in a big way in the past year—but before jumping in, there are a few important organizational questions to consider with your team.
1. What are the goals of monitoring and analysis?
Thought starters—goals may include:
- Measuring the results of a specific program
- Gaining feedback on traditional advertising
- Keeping an eye out for potential crises
- Integrating customer service to respond to issues
- Spotting opportunities for creative PR initiatives
- Keeping an eye on competitive initiatives
- Collecting data over time to add aid in market intelligence
- Gaining knowledge for new product development
- Finding out who is influencing the conversation
2. What cultural shifts will be necessary to get the most out of monitoring and analysis?
Thought starters—shifts to address may include:
- Will the customer service team need to be linked more closely to PR?
- Will the results of online analysis be tied into advertising creative or media buying decisions?
- Does it make sense for the team to meet on a regular basis to discuss opportunities and general consumer feedback from the research?
- What voice will be used to respond to questions and issues?
- If a crisis or issue does arise, how will a response be handled?
- If an issue arises on the weekend, who will take charge?
3. What tool(s) will we use to collect conversation data?
There are an overwhelming number of tools available for monitoring—the best tools will vary from one organization to another based on budget and specific goals.
4. Who will execute the monitoring and analysis?
Will the organization hire an internal team to manage monitoring and analysis? How will responsibilities (and budget) be split among brand teams, corporate teams, customer service and/or insights teams? Will the organization benefit from hiring a specialty firm?
5. How often will analysis reports be delivered?
Common reporting periods include:
- Daily monitoring with weekly analysis reports
- Daily monitoring with bi-weekly analysis reports
- Daily monitoring with monthly analysis reports
- Quarterly analysis reports
- Benchmark and follow-up analysis reports for specific programs
Before you spend a dime on monitoring and analysis, be sure to rally the team and use these questions as a guide to get everyone on the same page. The overall value of the monitoring program will be significantly higher.
Posted in Culture by Allie | No Comments
Take a closer look at what Loudpixel has to offer this Thursday, May 13 at TECH cocktail Chicago 13. Loudpixel will be demoing along with 11 other local startups and will be among the 500+ expected attendees at this year’s event. We will be demoing our internal reporting dashboard and will be on-hand to discuss all that our research and monitoring services have to offer.
Come out, grab a cocktail and find us in the start-up demo area to see for yourself how Loudpixel can help you make the right decisions for your company or brand on the web.
Event Details:
What: TECH cocktail Chicago 13
When: Thursday, May 13, 2010 from 6:30 PM – 9:00 PM (CT)
Where: John Barleycorn Wrigleyville, Chicago, IL
RSVP for a reduced ticket price ($10 online, $20 at the door).
We hope to see you there!
Posted in Events by Lesley | No Comments
Last Friday, Jeff and I were heading from Chicago to Ft. Lauderdale, Florida, to visit Ryan and catch up on some Loudpixel software development in person. When we arrived at Midway about an hour and 20 minutes before our flight, we realized, to our immediate horror, that our flight was out of O’Hare.
About 45 minutes and a few forehead slaps later, we arrived at the correct terminal in time to make our flight. To abate the embarrassment of arriving at the wrong airport, we decided to dive into Radian6 to find out how many others had made a similar mistake in the past 30 days.
137 people in the U.S. have told a story about going to the wrong airport in the past 30 days.
- 80 Twitter posts
- 51 blog posts
- 3 Facebook posts
- 2 Comments
- 1 Mainstream News articles
According to eMarketer, there are 26 million adult Twitter users (8.38% of the U.S. population, according to the U.S. Census Bureau). So extrapolating this out, let’s say that 926 people go to the wrong airport every month. Now we don’t look so bad, right?
And if that doesn’t make us feel better, here are a few more stories that have popped up in the past 30 days that make our arrival at the wrong airport pale in comparison.
Posted in Culture by Allie | 1 Comment
Update: This position has been filled. Thanks to all of those who took the time to apply!
Loudpixel is seeking a part time Online Research Assistant (15-25 flexible hours per week) to support the execution of online brand monitoring and market research.
Responsibilities include:
• Analysis of social media posts
• Maintenance of relevant keyword/search terms
• Analysis of broad trends in online conversations
• General marketing and new business support
Experience with Radian6, Filtrbox and/or general web development a plus.
Location
We’re based in Chicago and Ft. Lauderdale, but we don’t hold ourselves to one location—and we don’t expect you to either. We’re looking for a great candidate, whether he/she is in Chicago or remote (within the U.S.). As long as you have a computer and reliable Internet access, you’re good to go.
How to Apply
Submit cover letter and resume to Allie Osmar at allie@loudpixel.com.
About Loudpixel
Loudpixel is a Chicago-based firm that specializes in research-driven web application development and online monitoring/analysis. Learn more at loudpixel.com/about.
Posted in News by Allie | 1 Comment
When it comes to social media measurement and reporting, automated sentiment is often viewed as a miracle technology—one that saves time by replacing the need for any human intervention. But it’s important to look at this technology with a critical eye before accepting automated sentiment results as fact.
There are a few problems with the nature of natural language processing, which is behind automated sentiment. Even if a technology claims to truly understand the English language, it is still difficult for it to understand sarcasm, misspellings or slang—which are all prevalent across social media posts. I should also note that Biz360, one tool that tracks social media posts, pulled down it’s automated sentiment tracking for Twitter because there are not enough characters to give context for proper processing.
To show a firsthand look at how posts are being categorized through natural language processing, I took a quick look at posts related to “girl scout cookies” (a timely topic). The following are just a few examples of posts that ended up with the wrong categorization.
Marked as Negative:
Marked as Positive:
While the tools for conversation analysis are improving at a rapid pace, it’s important to remember that human analysis is still an important part of the research process.
Posted in Tools by Allie | 1 Comment
At Loudpixel, we like to take the time to look beyond brand-specific research and point our talents toward greater cultural trends. In this case, we thought it would be interesting to take a look at what people are giving up for Lent this year. So we analyzed a sample from 27,786 posts* from Wednesday, February 17 to see what people had to say.
The following are the top 10 sacrifices posted online during Ash Wednesday
- * Posts sampled from Twitter, Facebook, blogs, forums, videos and comments using Radian6:
- Undecided (looking for recommendations): 14.3%
- Junk Food: 10.9%
- Technology: 9.8%
- Negativity: 4.5%
- Alcohol: 4.2%
- Swearing: 3.5%
- Sex: 2.4%
- Meat: 2.1%
- Smoking: 1.3%
- Coffee: 1.1%
- Other: 45.9%
Some of our favorite sacrifices:
I’m giving up the 140th character for Lent. All tweets between now and Easter will only be 139 characters. That’s the kind of Catholic I am —@timhammer
I thought about giving up procrastination for #Lent but decided I’ll do it next year. —@joenobody
Posted in Culture by Allie | No Comments
