Odds are, if you work at an agency and manage a brand, you’re already using social analytics or listening tools to monitor your brand presence on the Web. At first, the marketing promises of “real-time, social monitoring” sounded alluring. However, the reality is that most social analytics today often fail to meet your (and your client’s) expectations. Why?
It’s because most social analytics aren’t technically designed to do what they are marketed to do. In fact, there are three common myths you should know about social analytics:
1. Analytics are real-time.
The first thing you probably noticed is that your results are far from real-time. Daily or weekly reports look okay, but often your analytics results lag by a few hours at best or a few days at worst. This is because many social analytics platforms are based on traditional search index models, which use some form of batch processing. This is compounded by delays in search indexes. For example, earlier this year, some users were reporting that Facebook Insights were delayed up to 4 days on average.
- Ask your social analytics provider what kind of lag or latency you can expect. Choose solutions that don’t rely on waiting for search indexes to refresh.
- Don’t be impressed if you see “live” Twitter or Facebook streams! The question is: do they do real-time analysis on those live feeds?
2. Bigger is better.
The most common question your clients will ask is: “Where do these results come from?” Most social analytics providers will say they have the biggest library of sources. However, bigger is not better in most cases. For example, imagine searching all of Google for your niche industry. You will get many results that are not relevant to you.
The bigger the library, the more noise you get – unless you can control exactly what goes in it. Otherwise, it’s hard to explain where (and in what context) you got your results.
- Avoid “black box” solutions.
- Look for solutions that let you control and curate all the sources you want to analyze.
3. The machine is always right.
While algorithms are always getting better, the reality is that no machine in the world can perfectly understand human sentiment. After all, that’s why your brand or client relies on your years of expertise!
The problem is nearly every social analytics provider claims they have the “best” algorithm. However, knowing that no algorithm is perfect, it is equally important to evaluate what human curation features there are, such as keyword training, error-checking, and team collaboration features.
- Check if your social analytics allow you to contribute and account for your team’s expert opinions.
- Know that human curation features also help demonstrate your expertise to clients and allow for “human override.”
- Demand real, real-time analysis (< 2 min refresh).
- Control your corpus & sources of information.
- Build best practices for team input.