The daily update report was generated at 7 a.m. It was reformatted by a second job, and then split into various versions by another, before going on a random walk between departments. It landed in inboxes and Slack channels. A new analyst joined and was given the job of generating the report. She was told it was vital that the report was generated as it was widely used throughout the business. The sky would fall in if it wasn’t received.
As part of her induction, the analyst had coffee with a trader who received a particularly complex version of the report. Keen to make an impression, she told him that she generated his daily internal briefing. She asked him what he did with this vital information. He had no clue what she was talking about.
Stunned, the analyst went back to her desk. Curious as to what would happen, she decided not to generate that version of the report the next day and waited for the sky to fall in. She’s still waiting.
She took a look at the other reports she was generating, and interviewed the users of the data and her peers in her department. As she examined the processes, she found that the path the data took through the business made no sense, with almost identical versions being created for different departments, and some versions spawning fresh summaries for others.
She found that all of the processes had started out quite differently, but had grown over time. New requests often asked for similar data to what was already generated, but each had a specific need. Each of these had been resolved by tweaking the existing processes. Nobody ever asked if the reports were still useful, or what was being done with the data. She started to redesign her processes so they made sense to the business as it was now.
The Sound of Silence
This happens a lot as organisations pivot and grow. The once-vital report becomes an ignored email that keeps getting sent. Another part of the business that has an issue needs data a bit like what a process already publishes, so a second report gets bolted on to the existing process. A once-simple Jira workflow has eleven steps, two of which are manual approvals that were put in as response to incidents. It makes sense at the time; it’s quicker than building something new. Then another bolt-on happens, and another, each a response to another reasonable request. At some point, the process becomes unrecognisable from when it started. Data’s being moved around the business, but the why has been lost.
Nobody reacts to a report nobody reads, so the process continues. Silence is how we tell most processes are working. As long as the process works, nobody complains, so the process continues. However, silence is also what happens when nobody’s using the data. The sounds of success and failure are the same.
When teams are run to maximum capacity, the logical thing to do is to bolt something on to what you already do. There’s no space allocated for stepping back and redesigning the process. The cost of stopping feels prohibitive. We reason that the bill will be paid by whoever owns the process in six months time. More often than not, that person is us.
Looking Sideways
Often, business processes are designed top-down, reflecting the organisation structure that supports them. This is the business-focused cousin of Conway’s Law, which states that systems mirror an organisation’s communication structure. Designing a process based on organisation structure can make for inefficient processes, where the data bounces around in ways that don’t make sense from the process’s perspective. When (re)designing a business process, it’s good to look at it side-on as well, following the data on its journey so you can understand how it can get where it’s needed faster and more easily. That’s what the analyst did.
As she reviewed the way in which her reports were used, she checked which ones were really still relevant. She was respecting Chesterton’s fence, the principle that you should not remove a rule or structure without understanding its purpose. There’s a tendency we all have when we encounter some random bureaucracy to wonder what the person introducing the rule was thinking. It’s a good idea to go and find out before you change it, in case there is a rationale that still exists but is no longer clear.
A useful question for teams to ask is “How do we protect our future selves from us?” It’s a nice way of ensuring that the cost of maintaining the fastest thing is considered. The team may still go ahead with the bolt-on, but agree a date at which they’ll come back and review the process. One of the things that helps is to set aside time for regular reviews of any bolt-ons that have happened. A schedule forces the conversation without waiting for a complaint.
This regular scheduled review doesn’t need to take long. It can be as short as a quick 15 minute review once a quarter. Do what the analyst did: ask your customers if they’re still using your output, if it’s valuable and what can improve. For any step that was created to stop something happening, find out what it stops, and if the risk is still real. Then, look at the journey the data goes on from the data’s perspective. Does it still make sense?
If you have a process that keeps growing legs, reserve some space to change it. Without a date in the diary, you’re just waiting for the sky to fall in.



