Quality as a process: How to create good product content

The challenge: Ensuring quality across large volumes of text

Writing quality copy takes care, time and experience - and the content team can assess the quality of a single piece of copy relatively easily. However, the perspective on text quality changes dramatically when it comes to creating and assessing the quality of a large number of texts, rather than just a single one. The available time resources become tighter and there is less room for careful drafting and critical revision. The quality of the texts will inevitably differ from that of a single, well-edited text. Nevertheless, certain quality standards remain essential. But how does an e-commerce company manage to ensure that these standards are maintained across such a large volume of text? 

Why a focus on process is crucial

Journalism researchers have developed interesting approaches to the question of the quality of their work, focusing less on evaluating the end result and more on the actions and processes of journalistic work. Put simply, the underlying thesis is that if the writing and editing process is carefully planned and executed, good texts will be produced. A well-thought-out process is therefore the key to producing consistent, high-quality content on a large scale.

Rather than focusing solely on the quality of individual texts, process orientation aims to ensure quality at every stage of production through clear guidelines, regulated processes and targeted reviews. In this way, a repeatable standard is created that can be maintained even when large volumes of text are produced.

The stages of a quality process in content production

Build Consistency: Setting Objectives and Creating Briefings

The foundation for quality content is laid long before the first word is written. A detailed briefing describes the objective of the content, the form and content requirements, and the structure. Such a document serves several functions in the content creation process:

• It assures that everyone involved knows exactly what the content should accomplish and how it should look.
• It ensures consistency and serves as a source of inspiration for authors and as a basis for automated creation with AI.
• It also serves as the basis for reviewers' feedback and automated quality assurance.

Possible contents of a briefing: 

Ensure accurate information: the research phase

An organized research phase will ensure that the information in the texts is complete and accurate. For similar products, a well-organized information gathering process eliminates duplication of effort, since data and information that has been researched once can be used for multiple texts. This saves time, reduces errors, and ensures consistent, high-quality content.

Data as the foundation: preparation is key

In e-commerce, the creation of product content is usually based on data from internal PIM systems, such as technical specifications, dimensions, or materials. This data forms the basis for accurate and complete text, but is often not directly usable for content creation. It must be prepared, for example, by consolidating, checking for completeness, and resolving inconsistencies. The type of data preparation depends on the chosen authoring approach - whether it is automated, manual or AI-based.

Knowledge Transfer: Getting Information from the Expert to the Content Writer

One of the challenges of content creation is that product knowledge is often held by internal experts, such as product managers or customer service. Knowledge exchange processes such as briefings or workshops help to transfer this knowledge into the written content. Alternatively, drafts can be reviewed by experts to check for accuracy and completeness. This helps to guarantee that the content is factually correct and targeted to the right audience.

Produce quality: Writing the content

There are several approaches to writing product content: The traditional way, in which a writer creates copy one at a time, provides accuracy but is time-consuming. AI text generators deliver large volumes quickly, but are often inaccurate when it comes to specific details. Data-based deterministic generation produces text according to fixed rules, providing consistent quality and speed, but with limited creativity and initial overhead.

The choice of method depends on the volume of text, production speed, and error tolerance. Automated generation using AI models alone is not suitable for security-relevant content, as it could invent facts. Here, the traditional approach or AI drafting with human review makes sense. For standardized products with clear data points, deterministic generation provides scalability and consistency. The key is to choose the approach that matches both the complexity of the product and the quality requirements in order to deliver high-quality texts that are appropriate for the target audience.

Assuring quality: Review and Feedback

Dedicated feedback loops for targeted quality control

In the review phase, specialized feedback loops are used to manage quality by separately reviewing aspects such as SEO, content accuracy, and style. This approach increases efficiency and improves error detection. The workflow can be flexibly adapted to the text group: complex content types pass through all reviews, while standardized product texts require simplified loops.

With large volumes of text, it is often not possible to check each text individually. For deterministic text generation, it is usually sufficient to check the set of rules and the interaction between data and rules, and to supplement this with the help of additional samples.
Samples are also a useful means of assessing the quality of AI-generated text. In addition, automated scoring systems based on defined quality criteria provide valuable insight into the overall quality of the text.

Automated support for targeted feedback loops

Automated tools enhance the review phase by supporting correction steps such as highlighting word choice, deviations from briefing specifications, or SEO violations. This dramatically speeds up the review process. In addition, these automation tools can process large volumes of text, identify patterns, and provide insights for process optimization.

Post-Publishing: Monitoring and KPIs

Quality assurance does not end with the publication of product content, but continues with monitoring. KPIs such as conversion rates and customer feedback are used to measure the impact of the copy. This makes it possible to check whether the content is achieving the desired goals. Based on this information, adjustments can be made and the content process can be continuously improved. This ensures long-term quality.

Additional tools

Training for content creators and editors, clear responsibilities, regular meetings (editorial conferences), and sufficient and up-to-date documentation of processes and specifications are important additional tools for success. This ensures that everyone is up to date and that consistent standards are maintained. Through continuous feedback and regular reporting, the process can be improved iteratively.

Improving process quality means ensuring content quality
As is clear, quality in product content production is a continuous process that must be constantly monitored and developed. A structured quality process offers many benefits, including scalability, consistency, and efficient use of resources - invaluable when dealing with large volumes of text. Each phase, from defining the requirements and research to content creation and quality control, plays a crucial role in guaranteeing and continuously improving text quality.