top of page
Writer's pictureVinay Payyapilly

Writing based KPIs for technical writers


I have written earlier about Measuring Effectiveness Through GAI-Driven Support Outcomes as a way to measure the performance of technical writers. Using the effect of content on the volume of support tickets aligns the technical writing team with the organization’s goals. Along with this, it is also important to measure the progress of a writer within the practice. As leaders, it is incumbent on us to push our teams to strive to become better at the practice of technical writing.  


In my team, I have implemented the concept of a review score. This is a simple numeric field on the ticket that the content reviewer increments by one every time they ask for a fix in the article associated with the ticket. Interestingly, this has brought about several changes in behavior: 

  • Writers use the tools available to them to check grammar before they send it for review. 

  • Writers learn from each review and carry it forward to current tasks. 


A common debate among technical writing leaders is whether we need editors in a TW team. The usual argument in favor of not having an editor is the availability of tools such as Grammarly or Word’s Editor. More recently, ChatGPT and other generative AI (GAI) tools have entered the conversation. These tools, they argue, makes the editor redundant by ensuring proper writing. But this argument misses a key point, and I am guessing it is because of the use of the term editor. If we go by the literal meaning of the term, I agree that we don’t really need editors in technical writing. However, if we were to change that to ask, “Do we need reviewers” then the conversation might be different. No, not a technical reviewer but a content reviewer.  


Every article on a Help site must answer three questions – what can you do, how do you do it, and why would you want to do it.  


Most of the time, we focus on the first two and pretty much get it right. The last question is something most of us miss. Whether the article answers the third question or not is the difference between a good article and a great one.  


Let me illustrate this with an example. Assume there is a setting with the label – Require mobile number. It is a check box that can be turned on or off. Simple enough to explain. The typical article heading would be – Make unique mobile number mandatory. The article would go on to tell the user that after they enable this setting, whenever they add a new customer’s details, they will have to enter a unique mobile number. Nothing wrong with this article per se. But think of everything we left out. We don’t tell them the benefits of enabling this setting to the business. For instance, it reduces the chances of having duplicate records which might lead to inflated expectations from marketing campaigns. It also allows the business to maintain an accurate history of all dealings with that customer.  


In a GAI-enabled world, including the why in our documentation takes on even more importance. A bot will not have the context that a human would and hence it is important to write more complete content. Key to this is setting the right incentives for the writers.  


Comments


bottom of page