Analysis of texts (app reviews, emails, chats, phone transcripts) is essential for proper product development planning. Clusterization of texts and extraction of keywords can help organise a backlog of projects, identify those with the biggest impact. How to make it less manual and optimised?
This discussion focuses on two methods of conducting text analysis for product development: ABC and XYZ analysis. By using these approaches, it is possible to gain a comprehensive understanding of the textual data and derive meaningful insights to guide the product development process.
The first step is to extract keywords. This problem is out of the article scope. Let’s assume that the problem of keyword extraction is solved using one of the Keyword and Keyphrase Retrieval techniques like KeyBERT. One of the approaches to make the analysis less manual is to use inventory management technics. Let’s focus on how to get analysis of texts for product development using two methods: ABC and XYZ analyses.
The ABC method is based on the Pareto principle – 80-20 rule. ABC analysis is a technique used to categorize inventory items based on their importance. The ABC classification goes like this:
- A items are important and need close attention
- B items are semi-important and require some attention
- C items are of low importance and can be given less attention
Importance is typically determined by considering factors such as cost, attrition and impact on quality. In this context, the importance of a keyword could be measured by the amount of time it takes support agents to resolve a customer issue. Categorising keywords according to their importance allows easier planning and prioritisation of feature requests. For example, if A keywords correspond to cases that are critical to the success of a customer or business, they should be prioritised above all others. Conversely, if C keywords represent issues of minimal importance, they can be given a lower priority. By using the ABC analysis method, teams can allocate their resources more effectively, thereby optimising their workflows and achieving better results.
Another valuable technique for inventory management is the XYZ analysis, which involves identifying items in a company’s inventory that have stable sales, high turnover, or are slow-moving. The analysis is so named for the three categories into which items are sorted:
- X items are those with the highest turnover
- Y items are those with moderate turnover
- Z items are those with the lowest turnover
In the context of support operations and keyword analysis, the goal is to gain an understanding of the frequency with which these items appear within a specific timeframe. Typically, this timeframe is set at weekly or monthly intervals, although businesses with a high volume of support requests may need to measure keyword frequency on a daily basis. By analyzing keyword frequency over time, organizations can gain valuable insights into the demand patterns for specific products or services.
While ABC and XYZ analyses can be valuable tools, they may not be sufficient when dealing with a high volume of extracted keyword combinations. In such cases, it can be helpful to combine the two methods and group the keywords into nine categories based on their importance-frequency combination: AX, AY, AZ, BX, BY, BZ, CX, CY, and CZ. The AX group is the most critical, as it contains items that are both important and frequent. These items should be given top priority when planning inventory and support operations. The remaining groups can be planned in any order, based on the specific needs and priorities of the organization.
When applying the ABC and XYZ analyses, it’s essential to consider the appropriate proportions for your specific business and customer base. While typical proportions for ABC are A=70%, B=10%, C=10%, and for XYZ, X=70%, Y=15%, Z=15%, it’s important to note that these proportions are not universal. Rather, they should be applied based on your knowledge of your customers and your business.
| X | Y | Z | |
|---|---|---|---|
| A | Important and frequent | Important | Important but rare |
| B | Important | ??? | ??? |
| C | Not important but frequent | ??? | Not Important |
In summary, the pipeline for prioritisation using a combination of ABC and XYZ analysis and importance-frequency categorisation is as follows:
- Keyword extraction: Use a tool such as KeyBERT to extract relevant keywords from customer inquiries and other sources.
- Definition of importance and association of it with keywords: Determine the importance of each keyword based on factors such as cost, turnover, and impact on quality. Categorize each keyword as A, B, or C, or X, Y, or Z, based on its importance and frequency.
- ABC analysis for keywords: Use the ABC method to categorize keywords based on their importance, assigning them to the appropriate group based on the proportion that best suits your business.
- XYZ analysis for keywords: Use the XYZ method to categorize keywords based on their stability, turnover, and speed of movement, assigning them to the appropriate group based on the proportion that best suits your business.
- Extraction of cases from AX and the rest of groups: Identify the most critical group, AX, and extract cases from this group for further review and action. Then, review cases from the remaining groups in any order based on their importance and frequency.
- Backlog review using cases from step 5: Review the backlog of inventory and support requests using the cases extracted in step 5. Prioritize requests based on their importance and frequency, and develop a plan for addressing each request in a timely and efficient manner.
Regularly reviewing your backlog is essential to ensure that it remains focused on the most important items for your customers. Implementing inventory management techniques can help to ensure that the backlog includes the most impactful potential items for development, reducing the need for manual work. After implementing ABC and XYZ analysis to categorize your keywords, you can consider other prioritization techniques such as Kano analysis, MoSCoW, or KJ method to further refine your backlog.