After a short hiatus from the series, I am proud to say we are back with another new business model. This week we are going to take a look at the data mining business model. Using this model means that revenue is generated as a result of data mining through the usage of a product. Nearly every business can use data mining as a way to increase revenue, but there is a growing trend to use data mining as a principle revenue stream. However, this approach does not come without its pitfalls.

Overview

Data Mining is the practice of collecting data from the users who interact with your product. This data can range from personal user information to basic usage statistics. Once the data is collected, it is then sold off to outside companies for profit. The data is then used primarily for marketing purposes, so the purchasing company can improve upon advertising techniques and product placement. Businesses like to use this model because it allows the company to offer their product for free and just sell the data. Although, businesses have to be careful about what type of data they chose to collect.

Selling a user’s personal information is generally a very bad business practice. It is required by law, in most areas, to publicly post a statement that tells the user that their personal information is being given to outside parties. If you’re not sure if your favorite applications are doing this, check out their privacy policy. Collecting and selling personal data can mean certain death for a business. If your user base catches wind that you’re selling their personal information, you will lose a marginal amount of users due to privacy concerns or begin collecting useless user data from those who refuse to enter real information to use your product. Both of these scenarios will end up causing a largely negative impact on your business model. So why would anyone in their right mind want to use data mining? Well, as I stated earlier, there are better options when it comes to the type of data your business chooses to sell.

The best type of data to sell to companies fits into two categories:

• Data which will save the buyer money by purchasing it from you
• Data that provides new insights unique to the usage of your product

Of these two categories, the first is more common. This is because the data tends to be quantitatively focused and generated organically through the use of the product. The selling point of this data is at the mercy of the buyer. They will not pay the asking price if it is above what it would cost for them to collect the data themselves (through surveys or other market research). I do not mean to imply that this data is not a valid option for a business model, it just means you must have a large quantity of data to sell or a large buyer market for the data to make the same amount of revenue as the second type.

If your data meets the second category, then you have hit the golden goose. This is the data that companies will pay a high premium to obtain. Using your program, you have come up with a way to pair the data between different components to provide your buyers with an unexplored perspective. The main hurdle now is to prove to the company that this data is useful to them. Proving the validity of data comparisons can be a quite difficult feat and should not be taken lightly. If you are going to sell this type of data to buyers, you better be able to show the impact this data can make on their business. Without showing this, your business will fail and should begin looking at a different model to drive revenue.

Advantages

• Low Maintenance
  • The data collection process is an automated process and only requires minimal interaction in order to package it for delivery.
• Market Size
  • The data can potentially be tailored for many different markets and can be sold to multiple businesses inside the same industry. Having a large potential market insures your data will always have a buyer.
• Large Revenue Stream
  • Depending on the quantity and type of your data there may be companies lining up to buy your data at a premium. This makes your business model very lucrative provided the data has a high market value.

Disadvantages

• Privacy
  • As mentioned earlier, the type of data you sell may cause users to boycott your product all together. If you can avoid selling personal data (or any user-specific data) than this is point does not apply.
• Validation
  • Coming up with the techniques to ensure the data your business collects is valid can be quite difficult. Users entering false information and automated bots/spiders/users are just two of the things that must be detected in order to ensure valid/marketable data.
• Marketing / Reliability
  • Creating market demand for data can be a difficult task. Once you have created the market it then shifts to maintaining the market. Making sure your data is relevant to be sold on a repeat basis means that longevity to your business. It is difficult to find this type of data.

Conclusion

As with any business model, data mining has no shortage of disadvantages. I personally am using data mining in an upcoming project and feel that in the right market it can easily surpass any of the alternatives. The problem lies in finding the data and discovering the potential market. If this seems obvious to you, I highly recommend giving data mining a try.