When you discover a bug in a flow with lots of steps, isolating & documenting it can take lots of time depending on the number of steps and the level of detail required to provide in JIRA.
Writing the steps of reproduction & test data in natural language, including screenshots & video, collecting network and console logs can take valuable time from testing (more than 30 minutes for a single issue) and it’s not the most fun activity once the bug was found.
When you test slight variations of the same long flow (e.g test a checkout process with a discount code, with payment via Paypal, with a discounted product, etc), the majority of the steps remain the same and we have to redo them over and over again for each variation.
When the flow is long (over 20 manual steps) and the number of options to test is high (over three) a lot of effort is required to repeat the same actions until we notice the actual result.
When the same flow generates different outcomes for different input data, redoing the flow several times can become time consuming when the flow is long and the number of test data grows.
For example when checking the cart total of a shopping cart we need to verify pricing based on where the products are being delivered and with the associated discount codes.
When you are validating the application against invalid sets of data in a signup, login or checkout form, the process can become trivial and tedious when lots of field values have to be checked.
Generating the set of data for each field is a process of creativity but executing each field test is not always fun when the number of fields grow or custom UI controls are involved or the process is multi-screen and you cannot use a form filler.
When you have to test a long flow every week because it's really important, it can become tedious to redo it if the flow it's hard to automate and there's no changes to it.