Build or buy? A software guide for London SMEs.

Map the work that happens outside the main system

Look for the spreadsheet beside the CRM, the client approval in a message thread and the weekly report assembled by hand. Those workarounds are clues, not proof that the whole system needs replacing. Ask why people leave the product: missing functionality, confusing configuration, poor data or an unclear process.

For a London interior design practice, accounting software may work well while client selections remain scattered across email. A small approval portal could address that gap without rebuilding invoicing, contacts and document storage. The useful scope sits between the existing tools.

Compare four options, not two

Build-versus-buy is often framed too narrowly. Configuration and integration deserve their own place in the decision. A subscription with the right setup may outperform a beautiful custom interface that nobody maintains. An integration can remove copying while preserving familiar screens.

Write one short option statement for each route. Explain what stays, what changes and who will own it. If the only argument for a new build is that it would feel more modern, return to the workflow and define a measurable improvement.

• Configure: use capabilities you already pay for.
• Buy: choose an established product for a shared business need.
• Integrate: connect agreed information between existing systems.
• Build: create the specific interface or behaviour the business is missing.

Custom software should solve the awkward part of the business, not recreate the parts that already work.

Choose websites and apps around their job

A public website, ecommerce store and private client portal have different requirements. A WordPress-led site can support an editorial workflow; a Shopify-led store can put product and order management at the centre. A custom application may be appropriate for a distinctive approval or operational process. Confirm the exact features and integrations before choosing.

Mobile app development needs a separate justification. Frequent repeat use, field work or device-specific capabilities can make an app useful. An occasional customer who only needs a project update may be better served by a responsive web portal without an installation step.

Include the unglamorous costs

Compare more than the subscription and initial quote. Data cleanup, migration, testing, access management, staff training and support all require time. Ask who updates integrations when a supplier changes an API, and how the business retrieves its information if the arrangement ends.

For a small London team, a system that relies on one person’s undocumented knowledge creates an operational risk. Request a clear account-ownership plan and practical documentation. A lower initial price is not necessarily the better option if routine edits or incident handling become dependent on an unavailable supplier.

Prototype the decision that carries the uncertainty

Do not prototype every screen. Test the part that could change the recommendation: can a client understand the selection view, can a coordinator resolve an exception, or can your current CRM provide the required information? Use representative examples rather than a perfectly clean demonstration record.

Give staff a task and observe whether they can complete it with the information shown. If the prototype exposes a policy disagreement, resolve that before development. Software cannot make an ambiguous approval process clear merely by placing it inside a new interface.

Define the exit and the first release together

A useful proposal states what the first release does, which systems remain authoritative and how success will be checked. It should also describe source access, data export and support responsibilities. These are practical ownership questions, not signs that a collaboration is expected to fail.

Begin with one team or customer journey. Keep the wider roadmap separate from the acceptance criteria for the first release. That gives the business a chance to learn before committing to a broader platform—and makes the eventual expansion a choice informed by use.

Put the idea to work.