Key takeaways
How to measure digital marketing ROI
Nobody has a perfect answer to this question. Not us, not the biggest agencies in the world, not the people who’ve spent years trying to build a model that actually works.
If that sounds different from what you usually hear from agencies, that’s the point.
To be clear about the claim: we’re not saying marketing ROI can never be shown. There are plenty of occasions where you can show the causal effect of marketing work. We’re talking about organic marketing, across SEO, GEO, ABM and digital PR. In these spaces it’s rarely straightforward, and the more channels you run together, the harder accuracy gets.
Our team has been wrestling with this for a long time. Like everybody in the industry, we’d love a clear calculator or formula we could share with prospects that effectively says: spend X, get Y out. So far, all we can honestly share is a few different ways of thinking about it.
What you’ll usually see is one of two extremes. One is the confident formula that doesn’t hold up to scrutiny. The other is giving up entirely because the data’s too messy. We believe the answer sits somewhere in the middle. If you’re comfortable with the numbers not being perfect, you can measure in ranges, accept being wrong by five to ten percent, and still have something genuinely useful.
This is written for anyone currently asking this question.
Why nobody’s worked this out yet
Measuring digital marketing ROI with complete accuracy is a well-known problem. We work with businesses across industrial, technology, health and wellness and professional services, and while the specific services, products and offers are different, the measurement challenge always ends up looking the same.
It isn’t a new problem either. One of the most famous marketing quotes is from over a century ago:
“Half the money I spend on advertising is wasted; the trouble is, I don’t know which half.”
– John Wanamaker (attributed)
A hundred years of technology later, the honest answer hasn’t moved as far as you’d hope. HubSpot’s 2026 State of Marketing report, with data from over 1,500 marketers, lists proving ROI as one of the top persistent challenges globally. That’s according to experienced marketers at established businesses with proper budgets and tools. Which confirms what we’ve already said: nobody really has a perfect answer.
Every corner of the industry has calculators and models promising they can finally prove the value. Sadly, a lot of them aren’t accurate. They’re convincing sales tactics with a nice interface. And even where genuinely good models exist, there’s so much noise to cut through that they get lost in it.
One of the core reasons nobody’s worked this out: the modern customer journey spans more areas than ever before. Organic search, social, email, trade press, word of mouth, events, direct conversations and everything in between. Tracking systems can only follow the parts of the journey they have visibility over, which is only a fraction of the whole.
As well as multi-platform journeys, we also have consent banners, cookie rejection, ad blockers and browser tracking prevention. The parts that are visible are shrinking. An experienced analytics expert told us recently that even with the best tracking setup in the world, you should only expect to see somewhere between 50 and 70% of your data. Think about that for a second. The dashboards most businesses treat as the full picture are, at best, a 50 to 70% sample.
And on an integrated organic project, running a few channels together, a lot of the work doesn’t even make it into that sample. Channels that work in the background, building authority, generating referrals through private conversations, building relationships, are a core part of what we do. It’s hard enough to show the value of the things that can be tracked, let alone the things that can’t.
Really, the only digital marketing channels suited to measuring ROI are paid ones. The platforms come with built-in reporting on conversions and revenue, because they’re deeply incentivised to keep selling ads. In our experience the numbers are often generous, favourable to the platform, but at least they exist. Organic has no equivalent unless you build it yourself, and even then you can’t capture all of it. Partly that’s because the journey has more steps. And it’s potentially also because the platforms have less of a vested interest in making organic’s value obvious: if it were laid out in black and white, plenty of businesses might spend less on ads. Every channel we work across, SEO, GEO, digital PR and ABM, is organic, so this is a problem we deal with every day.
The channels that don’t leave a footprint
Beyond in-person events and conversations, there’s also dark social. This is where a lot of content sharing, conversation and engagement happens, and it’s dark in the sense that it isn’t tracked or visible. Rightly so. None of us would want ad platforms or marketing agencies reading our private messages. But it’s worth remembering what sits there. WhatsApp groups, Slack channels, email forwards between colleagues. None of it is visible, and there’s a lot of unmeasured value in it.
What this might mean in practice for a client of ours: someone reads a piece of content, shares it with their procurement director in a WhatsApp message, and the director visits the website a couple of weeks later and converts. In Google Analytics, all we’d see is a direct visit. The content that started the conversation gets no credit at all.
As we’ve said, even with the best tracking setup in the world you can’t see everything, and forcing a revenue answer onto every output can undersell a channel. It can also affect the quality of the work if that’s the only thing you’re optimising for. Take digital PR: a strong backlink might not send any traffic directly, but it can still strengthen the website and the brand entity, and improve your chances of ranking for something six months from now that a prospect finds and converts through. It’s not always direct, and sometimes over-measurement can kill a good strategy. Same with coverage: a piece in the trade press that somebody actually reads and trusts mostly doesn’t leave a trackable footprint, but it builds confidence and makes the decision easier down the line.
AI search (GEO) has made this even harder
SEO, for all its complexity, at least comes with some data. Google Search Console shows you what people searched and which pages they landed on. It won’t connect any of that to enquiries, that’s not what it’s for, but configure your analytics platform and CRM on top and you can start tying organic traffic back to real outcomes. Not perfect, but workable.
Generative engine optimisation is harder to do this for, at least at this stage. We’re writing this in August 2026, so things may change. We’ve already started seeing AI search impressions appear in Search Console, and you can measure AI referral traffic in GA4, but it’s far from perfect. The bigger issue is behaviour. People aren’t typing one question and clicking through to a website any more. They’re having multi-step conversations, and a brand can appear at several points in that journey. The pages that show up early may never get the traffic they would have in traditional search, but that doesn’t make them worthless. If somebody eventually chooses you because of what they saw along the way, those pages contributed to the return. There’s just no accurate way to measure it.
Don’t just take it from us. A recent academic paper on measuring AI search visibility (visibility rather than ROI, which tells you where the field is) concluded that answers vary so much between runs that visibility has to be treated as a distribution over repeated checks, not a single snapshot. Visibility is real, but it’s hard to measure. The return is harder still.
So is there any point in trying to measure it?
We don’t blame anybody who steers clear of trying to work out ROI. It’s difficult, and it’s easy to get wrong. Agencies, freelancers, in-house teams: everybody’s in the same boat, so we get it, and we don’t judge it.
But from our side, we’re confident the work we do provides a return. So we’d be doing ourselves and our clients a disservice if we didn’t try to shine a light on the specifics.
There’s a harder truth here too. If you can’t present value, or at least show senior stakeholders, clients and decision makers that you’re actively working on it, you can lose the investment. It’s a harsh reality, but people in in-house roles lose jobs over this. Agencies lose clients. And businesses stop investing in something that’s actually working for them. We’ve seen it many times, especially with businesses that are hard to market: a manufacturing firm with a lead generation website isn’t confident in the total value of its SEO, so the budget gets withdrawn. Six months later sales are down and nobody connects the two.
Having that difficult conversation already?
If you’re struggling to show the value of your organic marketing, we’re happy to talk it through.
And to be fair, it’s a reasonable call from where the decision maker is sitting. If nobody’s shown them the value, why would they keep paying for it? That’s exactly why it’s worth agreeing a measurement framework early, one that everybody’s comfortable with. Realistic targets, honest caveats, and an open conversation about what the data can and can’t show. Do that up front and the work gets the time it needs to prove itself. Ignore the data because it’s scary, and all you’re doing is waiting for the difficult conversation to arrive.
Three ideas our team keeps coming back to
Fortitude is a collective of senior specialists. Everybody’s been in the industry a decent amount of time, everybody has tried to answer this question on many occasions, and everybody brings different ideas to the table. When we debate this internally, we rarely land on something everyone’s fully happy with, because it isn’t a clear-cut debate. But there are three ideas we’ve collectively agreed are worth exploring further, and we may even build calculators for them on the website in future (and no, we won’t pretend they’re perfect just to sell more work 😅).
One: working out what you can see and attribute directly. Two: time, efficiency and cost savings. Three: the returns that never touch the website but come as a result of the work. We keep them separate on purpose. Three rough views you understand beat one blended number you don’t.
Idea one: what you can see and attribute directly
Across the channels, this is the stuff you can actually see. Form fills, tracked calls, logged enquiries, and how many of those turned into customers, depending on the client at hand. Based on the number of customers that came through and the value of each customer, you can work out the return on the investment. How workable this is varies by channel: with SEO and ABM you can often count fairly bluntly, with GEO you sometimes can, and with digital PR you rarely can, for the reasons above.
As a rough example: if an average contract is worth £50,000 and a typical customer relationship runs to three contracts, one new customer is worth around £150,000 in lifetime revenue. Close one in four qualified enquiries and every enquiry carries an expected value of roughly £37,500. Apply the margin and you have a number you can defend.
Obviously this is a black and white, overly simplistic view, and it doesn’t paint the full picture. Which is exactly why we think it’s more honest to give a client three different views than one overall number.
Idea two: time, efficiency and cost savings
For a lot of our clients, before moving into content creation, outreach or more aggressive tactics, we help them revamp their processes to get more out of their time. That could be fully refurbishing their CRM, working with them on SEO knowledge and processes, or a number of other things.
Compare how efficient the team was before versus now. How many more deals pass through the system, how much faster the things happen that used to eat whole days. Based on their day rate, or what they charge their own clients, you can work out the cost you’ve saved them, and money saved is money made. As a rough example, a senior person on a £500 day rate getting five hours a week back is somewhere around £30,000 a year of recovered capacity. Account-based marketing sits partly in this idea too: because an ABM programme starts with a defined list of target companies and people, progress against that list is genuinely measurable over time, even before the deals land.
Idea three: the returns that never touch the website
One example to illustrate what we mean. We were on a Google Meet with one of our clients recently when a phone enquiry came in, so we paused the conversation while they took it. It turned out the caller had found them on Google and wanted to ask about a high-value commercial installation. They never filled in a form, and they didn’t appear in any of our tracking data. If we hadn’t happened to be on that call, and the caller hadn’t mentioned how they found the business, that enquiry would never have been connected to our work.
So something we encourage every client to do: whenever someone calls, emails or speaks to them at an event, ask how they first found out about the business. It definitely isn’t foolproof. But do it actively, mark the source against each lead in the CRM (‘found us online, enquired by phone’), and over a year you start to see how many of those enquiries became customers. Then, same as the first idea, you can work back from customer value versus spend to a rough return.
These are evolving conversations. We don’t think any of these methods are perfect, and we don’t know how long they’ll last. But they’re three measurement styles we think are interesting enough to try.
Want to try THESE METHODS on your numbers?
We can walk through what these three could look like for your business, and what a realistic return might be.
How should this influence budget conversations?
These methods don’t only need to be used at the tail end of a project, looking backwards. Once you’ve got one or two you’re comfortable testing, you can also flip them forwards: use them to forecast, and to have more honest budget conversations with prospects.
There are really two steps to it.
The first is forecasting a rough ROI. Take our three recent ideas, for example, and run them forwards instead of backwards: what could we realistically contribute in direct enquiries and customers? What could we save in time and efficiency? And what might come off the back of the work without ever touching the website? None of it will be perfect, but together it gives you something to work with.
The second step is the budget conversation, and it comes off the back of the first. Based on the return we think we can get you, and what a customer is worth to your business, here’s what a sensible budget looks like.
The benchmark we hear most across marketing is a return of five to one or above. Anything under two to one is usually treated as break-even once real costs are counted, and anything near ten to one is considered very strong. Benchmarks like these vary by sector and channel, so we take them with a pinch of salt, and you should too. But as a blanket sense check, it provides a helpful starting point.
The challenge is that budgets often aren’t set with any of this in mind. In many businesses, the marketing budget is whatever’s left after everything else, or a percentage of revenue borrowed from an old rule of thumb, or simply what it was last year. It’s rarely built from the question that actually matters: what are we trying to win, and what’s it worth to us?
If you’re chasing contracts worth £500,000, spending £1,500 a month on marketing means expecting a return of over 100 times your investment. Even the exceptional end of the benchmark is ten to one. That’s not to say outsized returns never happen, they do, but you get out what you put in, and it’s important to communicate that clearly at the beginning.
There are edge cases to all of this, which is a reason that the decision shouldn’t come down to the maths alone. We’ve had an electronics manufacturer return around 900% on a pretty modest budget, and we’ve won a security fencing supplier contracts worth considerably more than they were paying us. So don’t be put off if your numbers don’t fit the benchmark neatly. It’s a starting point for the conversation, not a filter on it.
Anyway, let’s round this up
Hopefully this has given you some food for thought, and maybe even something you can put to use. As we said at the start, there’s no one-size-fits-all answer, and we’re not going to pretend we’ve found one. What we can do, and have done here, is share our thoughts on a conversation the industry has been having since the dawn of marketing.
If you’d like to bounce ideas like these around, or get a clearer sense of how much value your marketing is really bringing you, book a discovery call with the team and we’ll see how we can help.