In many mid-market businesses a system landscape has grown over the years that nobody planned that way: a CRM in sales, project software in engineering, a task list in marketing, a shared drive for everything else, and mail as the actual operating system in between.
The reflex to treat this as untidiness is misleading. Each of these systems was introduced because a concrete need was unmet and nobody was responsible for solving it centrally. Sprawl is an ownership problem that shows up as a tooling problem.
What sprawl actually costs
- Duplicate entry: the same customer detail is maintained in two systems, and at some point they differ.
- Search time: before work begins, people search for which system holds the current state.
- Handovers: whoever passes work from one area to another copies by hand.
- Knowledge tied to individuals: only whoever introduced the system knows its quirks.
- Licences for systems hardly anyone uses any more, but nobody cancels.
Of these, the first two weigh heaviest, although discussions usually revolve around licence cost. Licences are visible, search time is not.
Clearing up without standstill
- 1.Record which systems are in use, who uses them and for what. Without judgement, otherwise systems get hidden.
- 2.Define a leading system per data type: where does the authoritative version of customer data, quotes, projects, documents sit.
- 3.Make duplicate entries visible and decide which are resolved by a connection and which by dropping a system.
- 4.Retire in small steps, each with a clear date from which the old system is no longer maintained.
- 5.Assign ownership: one person who agrees before a new system is added.
The last step is the only one that stops the landscape from looking the same again in two years. Without named ownership, clearing up is a one off exercise and not a solution.
The mistake that sinks clean up projects
The most common mistake is starting with the system that annoys most rather than the one that carries most. Remove a department's well functioning tool first because it does not fit the target architecture, and you lose the agreement of exactly the people you need for the rest.
The better entry point is a duplicate entry everyone suffers from. It is uncontroversial, affects several areas, and shows a result quickly that nobody has to defend.
We record the landscape and define with you which system leads per data type. That is the part that is hard to decide internally, because it touches responsibilities.