Operations people tend to be very liner in thinking, and tend to like to think of how to optimise every link on a sequence of a chain, and this can be a real super power for organisations transitioning from the entrepreneurial fire-from-the-hip style management of a small business into something investable and scalable, after all, without process, there is no automation, and therefore no scale.
But this superpower is a double edged sword, its possible to over engineer processes, or develop processes for lower priority aspects of the machine whilst ignoring more important parts. Its also possible to double down on processes that have already been defined, seeking out incremental 0.5% improvements instead of focusing on the elephants in the room.
At the beginning, it can also be a real challenge as so much need to be fixed, its hard to know where to begin… you can end out developing to process to develop processes, and end up forever in planning instead of making a difference to the business.
For the very initial stages when theres not much data or use cases, heuristic decisions should lead. Its better to have some kind of precedence and ‘rule’ even if it ends up getting changed in the long term over attempting to treat every instance as unique. The trick here is that any instance requiring a precedence setting decision is a data point you should be monitoring and looking to optimise in the future, these become the ‘decision points’ that form the foundations for processes, now they are identified, you need to conduct data collection to understand what the variables of that decision point.
As an applied example, we might look at a sales funnel and look at collecting leads, so we know that evaluating the quality of leads is a decision point, and what we need is data to better understand is the how to define differing levels of quality, so then this becomes the 2nd layer of focus. With data collection and discussions with the sales teams we might find that size of organisation, budget and number of employees are the data points that influence conversion the most, so now we have the variables on that decision point.
At this stage, we then need to consider the standards for the variables, again, you could spend a period of time to do data collection and come up with something more accurate, but at the initial stages, just pick something that makes heuristic sense and start from there. You can always adjust the parameters once you have sufficient data to do so.
So this gives you a relatively effective process to manage the various decision points in a business, stringing them together gives you a system with the data points to be optimised. Taking a step back, the question now moves to where to begin. In the absence of any data, just follow the funnel, its hard to optimise a stage in the funnel if there’s nothing hitting that stage of the funnel, so starting from the top and systematically working down is effective. If you have some data, then your first step is to look for bottlenecks, the areas where the largest (leakage) change in numbers is occurring and work on those first.
Ultimately the goal is to put something in place at the most obvious areas of leakage, us that to collect data, then after an incubation period of data collection, use the data to optimise that stage of the process.