Performance improvement is reflected through a comparison of the current state with the past. You can’t tell whether there is any improvement if you don’t have anything to compare with. You need a baseline for comparison purposes.
Sometimes it is difficult to establish the baseline when data is not readily available. If that’s the case, do an estimate. You can get a fairly reasonable estimate by doing the following.
What you do is to identify a scenario for a period of time when the condition is most representative for the work you are looking at. Write down the characteristics for that period of time. Then, collect sample data for a month. Note the differences between the month you collect the sample data and the representative scenario. Now, take the sample data and make adjustments based on your experience and knowledge about the business. The adjusted sample data is not perfect but would serve as a surrogate baseline. Apply a reasonableness test as a form of validation.
It is important to have a baseline for reference when you roll out changes to improve the status quo. Without it, it would be difficult to assess your effort and return on investment.