Tips and tricks for good product descriptions: How to optimize your product data

Good product data is crucial for sales. The days when customers relied on a single shopping channel are over. Today, they do their research online and investigate thoroughly before deciding to make a purchase. That's why it's so important to work with optimized product data. In e-commerce in particular, information must be meaningful, detailed, and always up to date.

Information that is crucial for success

Whether for product discoverability, information in a personalized newsletter, or an appealing product description—meaningful product and category information helps customers make decisions.

Product Data Marketing often involves processing and managing vast amounts of product information, which is used for campaigns and product desriptions and prepared in personalized messages for customers. This makes product data interesting for automation as well, because those who have a consistent product data feed with a good data structure and data quality can also use it for automated text generation.

However, not all product-related data that a company possesses is suitable for machine evaluation. In order for the data feed to be used for Content Automation, it must meet certain criteria. Consequently, the product information must be prepared accordingly in order to obtain appealing and meaningful product descriptions using automated text generation.

Product data and what it has to say

Let's first take a look at the data that most companies already have and that can be optimized. Often, existing master data from a PIM system or other product data management tool is well suited for automated text generation. Master data is simple, descriptive information such as product names, manufacturers, dimensions, etc.

A structured overview of this data (based on the model product = x, manufacturer = y, length = a, width = b, height = c) already allows simple sentences to be created that can even be used for all of a supplier's products. This gives us a short text such as:

The smartphone from x is available in size a x b x c.

Ultimately, it would of course make more sense to specify a screen diagonal here, but you understand the principle. Incidentally, the screen diagonal—or any other surface—could also be calculated automatically from the dimensions, opening up a whole new range of possibilities.

In the next step, we collect data on functions and features. Structuring this data requires a little more effort, and this is also where things often get stuck.

Product features as structured data

Most companies have a wealth of data on product functions and special features. This is important, because who would buy a smartphone based on the product description above?

But how can this be structured as machine-readable data? Since the assignability of data plays an essential role for a Text Robot, it is important to classify the product data into unique data fields. For example, it could say:

But also:

This product data can then be used to form a few more meaningful sentences:

The Android smartphone with touchscreen has an 8MP camera. With LTE data transmission, you can stay connected quickly and easily wherever you are.

The individual data fields can apply to multiple products and also contain data for completely different products in another data record. For example, the “operating system” data field for a desktop computer could contain the information “Windows 10.”

Meta data for interpretative analysis of information

So-called meta data, which enables a Text Robot to act interpretatively, is also particularly interesting for large e-commerce retailers. Examples of such meta data include sales figures, return rates, and reviews from other buyers.

Accordingly, a low return rate, for example, suggests that customers are very satisfied with the quality of the product, which enables the Text Robot to write a sentence from this data feed such as:

A particularly popular product that is guaranteed to give you long-lasting enjoyment.

Other interesting sources of information for Content Automation can also include shop information and manufacturer data, which offer additional value in automatically generated text and therefore enhance quality and optimization.

However, it is and remains important to structure the data optimally so that it is machine-readable. Fields that contain several poorly defined attributes are often unusable, and fields with continuous text cannot be easily evaluated by a Text Robot.

Product data optimization for automated content creation

But what does the optimal, purchase-relevant data for automated product description creation look like? If I want to prepare and optimize my product data for Content Automation, there are a few things to keep in mind. Specifically, this means that for successful product data optimization, I need to take a close look at my data and separate it if necessary.

In many cases, shop providers have, for example, a product description or product name that already contains the most important features. For example, the name of a sofa may already include the color, material, and perhaps even the dimensions. Even though this is practical for the header and allows the customer to see immediately what the product is all about, it is useless for a Text Robot.

Basically, such a data field contains a lot of wasted information – simply because the attributes from this data field cannot be read individually. However, if I have a separate data field for each property (“color,” “material,” or “dimensions”), I can generate a separate record for each attribute. When this is then played out in different variants and different sequences, variance is created and duplicate content is avoided.

The keyword here is granularity, and the more granular the data, the more flexible and multifaceted the text generation.

Frequency of occurrence or degree of completion

Of course, there is no point in creating hundreds of data fields that only apply to one product out of thousands. When it comes to Content Automation, it is important to weigh up the options: Which data describes most of my products and how complete will the data be if I create this or that attribute?

This line of thinking is not one-dimensional, because in some cases it can make sense to use an attribute with only a few entries for automated text generation—namely, when this attribute is a special feature that I want to highlight.

For example, it could say:

The iridescent bead appliqué makes this T-shirt a real eye-catcher.

Even if there are only a few T-shirts with this type of appliqué in my shop, this category is particularly important to me.

Granularity and structure

Start by creating 20-25 attributes and increase the number to a rule of thumb of about 90 attributes for the entire text generation project, which will already allow you to create a good variety. The rule here is: the more data fields I have with good content, the more flexible and detailed the automated text I can generate will be.

It may be a bit time-consuming to prepare the existing product data and product information for automated text generation by a Text Robot, but it's worth it! With well-structured and optimized product data, you can create great search engine-optimized product descriptions for all channels that appeal to and inspire customers.