Digital transformation for a mid-size company: a step-by-step plan

For a mid-size company, digital transformation is not one big programme with a single launch date. It is the steady replacement of manual work and disconnected tools with connected systems, one process at a time, each with an owner and a number that should improve. The plan below suits companies that are too big to run on spreadsheets but too small for a large internal IT department.

Step 1: Find where work gets stuck

Start with how work actually moves, not with technology. Spend time with the teams that handle orders, customers, finance and operations, and look for these signs:

• People type the same information into more than one system.
• Approvals happen by email, and someone has to chase them.
• Managers wait for a weekly or end-of-day report to know what is going on.
• Important knowledge lives in one person's spreadsheet or inbox.
• Customers ask for status updates because they can't see progress themselves.

Write each one down with a rough cost: hours per week, errors per month or days of delay. Rough is fine. You need enough to compare problems, not an audit.

Step 2: Pick one or two processes to start with

Choose processes that are painful, frequent and contained. A good first candidate involves one or two teams, runs many times a week and has a cost you can measure. Avoid starting with the process everyone argues about. You want an early, visible result that builds trust for the next step.

Set the target in plain numbers before any work begins, such as time to process an order, days to close the month or hours spent each week on reporting.

Step 3: Fix the process before the software

Automating a messy process just gives you a faster mess. Map the current steps, remove the ones that only exist because of old tools, and agree who owns each decision. Ask the people who do the work what they would change. They usually know already.

Step 4: Decide whether to buy, configure or build

• Buy when a standard product handles the process well, as with accounting, payroll or a standard CRM.
• Configure when a tool you already pay for can do the job with workflows, forms or integrations you haven't set up yet. This is often the cheapest win.
• Build when the process is specific to how you work, or when the real problem is that your existing systems don't share data.

Many mid-size companies end up with a mix: bought systems for common work, plus a custom layer that connects them. Our post on custom software development covers the build-or-buy decision in more detail.

Step 5: Get the data in order

Connected systems need agreed data. Decide which system is the source of truth for customers, products, prices and orders. Clean up duplicates before migration, not after. Agree names and formats once, so reports from different teams match.

Step 6: Deliver in small releases, starting with a pilot team

Release to one team or one site first. Let them use it for real work, collect their feedback and fix what gets in the way. Then roll out to the next group. Small releases keep the risk low and give you working results to show while the rest of the plan continues.

A wholesale distributor we worked with took this route with order and invoice processing. Orders arrived by email, in spreadsheets and as customer documents, and staff copied the details into the ERP by hand. We built one workflow that captures each order, checks it against business rules, updates the ERP once it is approved and sends exceptions to the right person. People still make the calls on pricing and approvals. The repetitive steps are automated.

Step 7: Train people and give the change an owner

New systems fail when people quietly go back to the old way. Give each process a business owner, not just an IT contact. Pick a few champions in the pilot team who can answer questions. Train with real examples from people's own work. Once the new route works, switch off the old one, so there aren't two ways of doing the same job.

Step 8: Measure, then choose the next process

Compare the results with the targets you set in step 2, and share them with the whole company, including what didn't work. Then go back to the list from step 1 and pick the next process. Repeated over time, these small changes add up to a different way of running the business.

Who you need involved

• An executive sponsor who can make decisions across teams and protect the budget.
• A process owner for each process, from the business side.
• A delivery team with product, design, engineering and QA skills, in-house or from a partner.
• Technical direction, so each new system fits with the last. If you don't have a CTO, a fractional CTO can set the roadmap and review the bigger decisions.

Why transformation programmes stall

• Buying a large platform before anyone understands the processes it has to support.
• Trying to change every department at once.
• Having no measurable target, so nobody can tell whether it worked.
• Leaving training and change management until after launch.
• Treating it as an IT project rather than a change in how the business works.

Where to start

Spend a week listing where work gets stuck, with a rough cost for each item. Pick the one with the clearest cost and the fewest teams involved. If you want help scoping it, our intelligent process automation team works on exactly this kind of workflow. If your plan includes moving older systems to the cloud, read how to estimate cloud modernisation ROI before you commit a budget.