A recent compilation of studies done by LTU Technologies has found a trend that is very concerning to people throughout the marketing and business world. On twitter today about 36% of all links that are shared go to an image of some sort, but since data mining programs only see the text and tags that are used it is impossible for them to identify what it is that the picture is of. Most reporting mechanisms can’t tell the difference between an image of your pet dog and the logo of a major company in an image.
This is a big concern because when companies use the information gathered to lean when, where and in what context people are talking about them. If, for example, a company finds that people in Dallas, TX are talking about their company in a negative way after a marketing campaign they know they need to adjust their strategy. There are, of course, many other things that companies can use this data for but when it is all in images they are left in the dark.
On sites like twitter the images are often tagged using the hash tag format (#hashtag) which has become the standard, but it is estimated that when brand name logos are used in the picture there is no hash tag about the company itself 77% of the time. This is a huge amount of data which companies are missing out on causing them to have an incomplete picture to work with when planning marketing strategies.
With millions of images being shared every day this is increasingly becoming an issue for companies who want to stay ahead of their competition and in line with their customer’s desires. The study done has shown that depending on the specific industry people commonly don’t mention the exact brand they are using in an image. There are, of course, also concerns with copyright violations that need to be investigated.
Some specific statistics found in the study include the following:
- Images of Beer only list the brand name 35% of the time
- Visual images are processed up to 60,000 times faster than text alone
- Pictures of food make up 20% of the images linked to on Twitter
- Pictures of objects make up 36%
It is easy to see why many marketing professionals working for big companies are concerned about this growing trend. As more people begin using Twitter and other social networks and specifically using them to post pictures it will only result in a bigger gap in the knowledge they have available regarding their customers.
Unfortunately image recognition programs are not efficient or accurate enough to sufficiently capture this type of information. Companies and information mining firms will have to work together to attempt to find a way to find a way to accurately measure how much people are really talking about their products because in today’s world people are communicating more and more through images and less and less through text.
