The phrase “digital transformation” implies demolition: rip out what you have, install the new world, endure the transition. For an enterprise that may occasionally be true. For a growing business it’s mostly a way to spend a year and a lot of money standing still. The better path, almost always, is incremental: transform the way your marketing and sales work one piece at a time, using as much of what you already have as possible.
The evidence favours the incremental route. The single biggest obstacle UK businesses report in marketing automation projects isn’t the technology; it’s integrating data across systems, cited by 67% as their main challenge. Big-bang replacements multiply that problem, because everything changes at once and every integration breaks together. Meanwhile the return on getting it right is substantial, with UK data attributing a roughly 32% marketing ROI uplift to well-implemented automation. The Value of Automation Report 2025 – DMA The prize is real; the demolition is optional.
Start with an honest audit of what you have
Most businesses own more capability than they use. The email platform has automation features nobody switched on. The CRM, if there is one, has fields nobody fills and workflows nobody built. The accounting system exports data nobody looks at. Before buying anything, map what you already pay for and what fraction of it you actually use. It’s common to find that the first quarter of “transformation” is simply using the existing stack properly, at zero additional licence cost.
Fix the connections before the components
The typical small-business stack isn’t bad; it’s disconnected. The website doesn’t talk to the email platform, which doesn’t talk to the sales pipeline, so humans re-key data and things fall through gaps. Connecting existing tools, natively or through simple middleware, often delivers more value than replacing any of them, because the failure was never in the components. It was in the gaps between them. This is unglamorous work, and it’s where the 67% integration problem gets solved or entrenched.
Replace only what’s genuinely holding you back
Sometimes a component does need to go: the email tool with no automation, the spreadsheet pretending to be a CRM past the point of coping, the system nobody will ever adopt because it fights how the team works. The test is specific: does this tool block a process you’ve already designed and proven manually? If yes, replace it, and only it. If the process itself is unproven, replacing the tool just gives your confusion better software.
The sequence, prove then expand
The working rhythm looks like this. Pick the process that leaks most, usually lead follow-up. Fix it manually first so you know what good looks like. Automate that one process on the smallest adequate tooling. Measure for an agreed period of time. Then take the next process. Each step is cheap, reversible and provable, which is precisely what a big bang is not. The fuller argument for the one-workflow start is in Where to Start with Marketing Automation.
A worked example of the incremental path
To make it concrete: a services business with a website, a mailing list on a basic email tool and sales tracked in a spreadsheet. Quarter one: enquiries from the website flow automatically into a simple CRM pipeline, ending the copy-paste and the lost leads. Quarter two: a welcome and follow-up sequence switches on in the email platform they already pay for. Quarter three: the CRM and email tool are connected so sales can see who’s engaging. Total new spend is modest, nothing was ripped out, and by year end the revenue function runs on a system. That’s transformation; it just doesn’t look like the brochure version. Should the processes bed in easily and the results be clear, then you would obviously not need to wait a 3 months for each step!
What to do yourself, and where help earns its fee
Plenty of this is genuinely DIY: the capability audit, switching on unused features, basic workflow building on modern platforms. Where external help earns its fee is the architecture, the data model, the integration design, the decisions that are expensive to unwind, and the discipline of an outside eye that stops the project sprawling. A few days of senior input at the design stage routinely saves months of drift later.
Transformation without demolition is slower to describe and faster to deliver. If you’d like an honest read on which parts of your stack to keep, connect or retire, that’s the conversation Digital Transformation to start with, and if the answer is “keep most of it” then that is what you will hear; we have not contractual obligations to any software providers or tech companies.
