Data Mining on a Budget: How to Use it for Your Online Shop

When talking about data mining, we refer to a special strategy of analysis that can be pretty time-consuming. However, implementation comes with many benefits, making data mining one of the most rewarding strategies to use in e-commerce. So, what is data mining actually about? How can you make the most out of it for your own online shop even if you only run a small company? What do you need to consider? Keep reading and we’ll give you the answers to all of these questions.

Index of this article:

⛏️ What is data mining?

❓ What is the purpose of data mining?

💡Example of data mining

⚙ Why use data mining for your business?

🧰 How to get started with data mining

🛠 Tools to use for data mining

👎 What to consider when using data mining

⛏️What is data mining?

When talking about data mining, we refer to the approach of going through data without having any particular focus or goal in mind. The aim is to discover elements that will help you improve your business strategy.

With data mining, you could, for example, look for any correlation between the different products that your customers buy in order to help you cross-sell better.

❓ What is the purpose of data mining?

When data mining, you begin your analysis without having defined any particular problem in advance.

You don't know what you're going to find or whether you're actually going to discover anything at all.

Usually, when conducting a research you would, for example, try to find out why your online shop has fewer website visitors on Saturdays.

Recommended reading:
How to Create a Survey for Your Customers

In other words: you would look for a particular information or data set. Data mining on the other hand, is basically about finding answers to questions you didn't know you even had.

The purpose of data mining is, therefore, less about discovering the answer to a particular question, but about finding useful patterns in your data.

💡Example of data mining

Referring to the book “Creating Value with Big Data Analysis” (by Verhoef, Kooge and Walk), the authors give an example of the British supermarket chain Tesco.

Tesco analysed its own data, looking for purchases made with the Tesco Club Card and found that customers that bought diapers tend to add beer to their shopping cart as well.

Tesco’s analysts also uncovered the best-selling items on Friday nights: beer and crisps.

Note: It’s not clear whether the company in our example actually was Tesco, as this example can be found in other sources referring to the American supermarket chain Walmart instead. This example is mainly to give you a broad idea of what you can do with data mining.

⚙️Why use data mining for your business?

Depending on the information you find when data mining, there are different ways of how to use it for your business.

However, an important benefit is an enhanced and more targeted promotion. Let's take the example of diapers and beer: if you were to sell both items in your online shop, you could use the information in a subtle, but smart way and place an offer or pop-up for beer on the diapers page (and the other way around), for example.

It is a fact that many customers preferably shop online during the weekend. Therefore, at the moment, most of your orders are made during this time, meaning that lots of parcels need to be shipped at the same time.

If you want to balance this rush in logistics, you could, for example, offer special promotions during the week for products that are the most popular during the weekend.

Make sure to announce and promote your sales in advance (e.g. on various social media platforms and in your newsletter). This way, your customers might wait a few days before making their purchase, thereby ensuring it to be less stressful for you to process orders in the future.

🧰How to get started with data mining

Shutterstock/a-image

You just learned about data mining. Now, it’s time for you to learn how to get started with it in order to get the most out of your research.

Unfortunately, the thing with data mining is that it takes a lot of time, especially if you want to do it manually.

We recommend you to go through your data step by step. For example, when focusing on products, consider all customers that have purchased more than one product in your online shop.

Which product is the most popular? Which products do customers that buy more than five products place in their shopping card?

You could also focus on specific product categories: if a customer purchased an item from the toy category, which other product category do they turn to?

Make sure to also have a look at preferences and correlations during different times of the day. Which products are particularly popular around noon, which ones around the evening?

Instead of focusing on your products, you could also take into account the different web pages on your website: which pages are the most popular at what time of the day? Compare your findings to your sales. Is there any correlation? This information can help you with your marketing campaigns and bidding strategies.

Good to know: there are a couple of useful tools for data mining. This way, you don’t have to do the analysis manually. However, many of them are rather expensive. Take your time and decide whether you'd rather spend your time or your money.

🛠 Tools to use for data mining

Of course, you can simply put all the data you can find in an Excel file and analyse it yourself, but it might be easier (and less time-consuming) to use special data mining tools instead.

Many providers of data mining tools charge hundreds of euros per month. Surely, as a small business owner, this would overrun your budget.

Instead, let us introduce you to a couple of tools that aren’t expensive, but affordable or even free of charge. This will allow you to start with data mining – on a budget.

Most tools offer a free test phase anyway, giving you the opportunity to even try different ones. Oracle offers a 30-day free trail for their data mining tool, for example.

Orange, on the other hand, is a 100% free, open-source tool.

👎 What to consider when using data mining

The process of data mining, as well as the outcome is unpredictable. Sometimes the findings are difficult to interpret.

Data mining can be time-consuming until you finally identify a pattern or promising correlation.

In addition, you should know that even if you discover a similarity in data, it doesn’t necessarily mean that one element influences the other. Sometimes there is no explanation for a correlation.

Here's an example:

On his website, Tyler Vigen has a whole list of data sets that seem to follow the same pattern, but end up having no real connection at all. Take a look at the graph below.

Source: tylervigen.com/spurious-correlations

We can see that the number of divorces in the US state Maine seems to be correlated to the per capita consumption of margarine. Could you therefore argue that only divorcees in Maine eat margarine?

Of course, there isn’t a real correlation between those two data sets.

That is why you need to be careful how you interpret your findings!

As you go through your data, consider what discounts you may have offered at the time or whether you offered a better price than your competitors during a particular time.

Make sure to also consider external factors such as the coronavirus: if you suddenly see an increase in board games, does this have to do with your actions, or with the coronavirus? Maybe both?

🔄 In short

Data mining can provide useful and even surprising information, that will certainly be of use for you and your business.

You might have to try a couple of different strategies or tools in order for you to find the most efficient way for you to analyse your data. In addition, make sure that you process the answers in the right way and that you don’t jump to any conclusions too soon.

The best thing about data mining is: you don't have a specific problem you want to solve, so you have nothing to lose!