We often find discussions, particularly on news sites, where keywords only appear once. Further related comments are not found by the tool and have to be reworked by a human being.
Obstacles for the crawlers
Further limits of the tools are pages like “xing.com”, which do not even let the crawlers in in first place. These pages can therefore not be monitored and miss in the evaluation. This is especially unfavourable for B2B companies. The same occurs for pages that are protected and have to be activated by clicking the subsection.
Spelling of the keywords
Of course the most common misspellings can be included in the search query, but this also leads to a higher rate of “lame ducks”. Mistakes can only be covered partly, as it is simply impossible to take every misspelling into consideration.
Evaluating the sentiment
Furthermore, the sentiment assessment of the tools is in most cases not accurate. They cannot evaluate ironic and ambivalent posts correctly. For example, a customer gives a product a good rating, but also criticizes the service. Whilst using the tool one has to tie himself down to one sentiment, which results in a loss of information. For that reason we do not decide on the base of the find, but on the topic and content of the post. This enables more specific analyses, which is especially important for companies with only 1.000 or less finds per month, as wrong evaluations have an even stronger effect on these.
Here is a further essential aspect of monitoring tools: SEO- and social media marketing measures of affiliates. An affiliate partner will always present a product or a service in the best possible way, as he is paid for every contract conclusion. This results in a positively distorted image of all articles, forum posts, tweets and Facebook-posts by this author group. Therefore, a machine-made evaluation (e.g. competitor monitoring) will probably lead to incorrect results. We have observed this at direct banks. According to a satisfaction survey of an agency, one of these banks was rated very positively, whereas the opinions of the customers said the opposite. This bank had thousands of affiliate partners who, of course, only advertised positive reviews on the web.
All these points prove that the quality of social media monitoring does not result from the tools alone, but from the operator. And as every analysis is only as good as the data it is based on, the social media monitoring should be performed by real specialists.