Decision guide

Website, workflow or software: what actually needs fixing?

A business can feel digitally broken in several places at once. The expensive mistake is to commission the most visible solution before locating the actual failure.

“We need a new website” often means enquiries are weak. “We need a CRM” may mean nobody owns the hand-off. “We need an app” can mean a service is hard to deliver consistently. Each proposed solution may be reasonable, but none describes the problem.

Before buying or building anything, trace one real customer outcome from first intent to completed service. Mark where confidence, information or momentum disappears. That usually places the primary failure in one of three zones.

The three failure zones

1. The market cannot understand or trust the offer

This is the outside of the business: positioning, proof, website and enquiry journey. Signals include the wrong enquiries, repeated basic questions, low conversion despite relevant traffic, or prospects who choose a clearer competitor.

A website change is justified when the service is sound but the digital representation makes it difficult to understand, compare or act. The intervention may be new positioning and content rather than a complete rebuild.

2. The team cannot deliver the promise cleanly

This is the operational middle: forms, queues, hand-offs, decisions and reporting. Customers may arrive confidently but then wait, repeat information or receive inconsistent updates. Staff compensate with inbox searches, spreadsheets and private knowledge.

The answer is usually workflow redesign first. Software may implement the new workflow, but automating the old one simply makes the confusion move faster.

3. The value genuinely requires a product

Software is appropriate when the customer or team needs an interactive capability that a clearer page and better process cannot provide: ongoing self-service, complex calculation, shared case state, high-volume structured decisions or a novel digital experience.

It should still begin with the smallest uncertain part. A working prototype can prove behaviour and operating assumptions before a full product absorbs the budget.

The useful distinctionA website changes what people understand and do. A workflow changes how work moves. Software creates or scales a capability. Some problems need all three, but the order matters.

A six-question diagnostic

  1. Where is the first observable failure? Before enquiry, during conversion, after hand-off or during delivery?
  2. Who experiences it? A prospective customer, a paying customer, frontline staff or management?
  3. What do people do to compensate? Call, chase, copy data, maintain a parallel spreadsheet, or ask the person who “knows how it works”?
  4. Does the desired process already exist? If two good operators would handle it differently, automation is premature.
  5. What would success change? Conversion, cycle time, error rate, capacity, customer confidence or management visibility?
  6. Can the assumption be tested without a platform? Better content, a manual concierge version or a lightweight tool may answer the important question.
EvidenceLikely first move
Good traffic, weak enquiries, frequent “do you offer…?” questionsPositioning and website journey
Strong demand, slow delivery, repeated chasing and rekeyingWorkflow mapping and redesign
Clear workflow, growing volume, manual rule-based workAutomation or internal tool
New customer capability with uncertain behaviourPrototype before product build

Choose the sequence, not just the solution

A clinic might need clearer service pages first, then a better intake hand-off. A fleet business might need the operational workflow mapped before an internal dashboard. A health-tech company might need to prove the clinical service manually before deciding what belongs in software.

The best first project is the smallest intervention that changes what the business knows or how the work performs. It should produce evidence for the next decision—not force the organisation to defend a large purchase already made.

Start hereBring one real customer journey, the current tools and three recent examples where it went wrong. That is enough to find the first useful move.