Key Performance Indicators or KPI for short is the indicators that drives most dashboards, or at least they should. These indicators are based on different values that can come from various dimensions or facts in the company datawarehouse. KPIs can be a mean the world to a business, but the business have to have the ability to react upon shifts in the indicators, otherwise the value o these indicators and flashy dashboards is deevaluated.
Look at a company that manufactures gearboxes for motorbikes, if they have an indicator that shows that the delivery of certain parts for their gearbox is to be delayed, they have to react to that scenario. If they are not able to, lets say get the part from an alternative sub-contractor then their productionline would be at a halt. On the other hand if they are able to get the parts from another sub contractor prior to the parts being out of stock, then the indicator have showned its value.
In the same company, they may have an indicator showing number of orders set up against the number of employees at hand for a given time. This indicator should take absence from work into consideration, this way management would have a chance to plan for the future. If for example there is an indication that orders is not flowing as fast as they could, and there is a lot of employees on vacation, then there is no need for finding extra workforce, on the other hand, if orders are flowing in at a higher rate than normally during summertime, where many employees will be on vacation, then the indicator would show a need for extra workforce.
KPIs is based on how the data looks right now in the datawarehouse, but they could be combined into what is known as predictive analysis. This is done on a dataset from the datawarehouse, ie. order data and employee absence from the last 3-4 years. If a predictive analysis is done on that dataset, and the algorithem is done properly, it could show the need for either more or less workforce during holiday seasons.
As with everything, visualization is of the outmost importance. Dont go over the top and try to show each and every KPI that your business could come to think of, on the main dashboard. Find a few 5-6 that defines the specific part of the business at hand, because to much information just tends to make the dashboard useless.
Here is an example that I think has way too much information, but that is just my opinion.
And as the opposite, here I have found an example of a dashoard that contains a few but important indicators.
I think that the latter is way more informative, and gives a fast oversight of the data, whereas the first gives me more information than I needed, and the oversight that is wanted is somehow lost.