The ROI of Google Business Profile Optimisation
Home Blog The ROI of Google Business Profile Optimisation: How to Calculate...
Blog · Tools & Agencies

The ROI of Google Business Profile Optimisation: How to Calculate and Prove It in 2026

How to calculate the real return on GBP investment - the metrics that proxy revenue, how to build a cost-per-lead model, and how to present ROI to clients.

Ampli5 Pulse Editorial Team April 22, 2026 8 min read Updated 2026

Every agency owner who has managed local SEO for more than a year has had the same conversation. The client asks what they got for the money, you open a dashboard full of views and clicks, and the room goes quiet because none of those words are money. The gap between profile activity and revenue is real, but it is bridgeable, and the bridge is a model you agree with the client in advance rather than a number you produce at renewal time.

This is how to build that model, defend the assumptions inside it, and present it to someone who does not believe you yet.

Why GBP ROI is awkward to calculate

Four structural problems make this harder than paid media reporting, and it is worth naming all four with a client before you show a single figure.

  • The conversion happens off-platform - someone taps call, has a two-minute conversation and books a job. Google sees a tap. It never sees the booking, the quote or the cancellation.
  • Google's metrics are interactions, not leads - a direction request is a person considering a visit. It is evidence of intent, not proof of a customer, and treating the two as identical is how ROI models lose credibility in one meeting.
  • The work is entangled - profile optimisation, website content, reviews and word of mouth all move together. Isolating the profile's contribution is an estimate, always.
  • The counterfactual is unknowable - some proportion of those callers would have found the business anyway. You cannot run the version of the month where you did nothing.

None of this means ROI cannot be reported. It means the honest output is a defensible range built on stated assumptions, not a precise figure delivered with a straight face.

The actions that proxy revenue

Four profile actions have a plausible line to revenue. Everything else is context.

  • Calls - the strongest signal in phone-led categories such as trades, legal and clinics. Someone dialling has largely decided.
  • Direction requests - the strongest signal in premises-led categories such as restaurants, retail and garages. Weak for service-area businesses, where it often means a supplier or a lost driver.
  • Website clicks - intent handed to your site, where it either converts or does not. Worth counting only alongside the site's own conversion rate.
  • Bookings and messages - the cleanest of the four, because the action and the enquiry are the same event.
Important
Profile views and search impressions are not revenue proxies. They tell you about visibility, which is useful diagnostically when actions fall, but a client who has been trained to celebrate impressions will eventually ask what those impressions bought. Keep them in the report as context and never as the headline.

Building a cost-per-enquiry model

The model has three moving parts: how many actions occurred, what proportion of each action type counts as an enquiry, and what the work cost. Agree the middle part with the client before the first report, because that is where every argument happens.

The table below is a worked example with illustrative figures, not results from any account. Numbers are shown in a single unnamed currency so you can substitute your own.

InputExample figureSource
Calls from the profile120Performance insights
Calls lasting over 30 seconds72 (60%)Call tracking, or a sampled week
Direction requests90Performance insights
Directions counted as enquiries18 (20%)Agreed discount rate
Website clicks from the profile240Performance insights, checked against analytics
Enquiries from those clicks12 (5%)Site conversion rate
Attributed enquiries102Sum of the three
Monthly cost (fee plus software)600Your invoice
Cost per enquiry5.88Cost divided by enquiries

Carry it one step further and it becomes a sentence a business owner can act on. Say the client closes one enquiry in four, at an average job value of 400. Those 102 enquiries become roughly 25 jobs and 10,000 of revenue against 600 of cost. Again, illustrative: substitute the client's real close rate and job value, and if they do not know either, that conversation is more valuable to them than the report is.

Set the discount rates once
The 60%, 20% and 5% above are the assumptions, and they should be conservative, written into the proposal, and left alone. A model whose assumptions improve in the months where results were poor is not a model, and clients notice faster than you would like.

Attributing a phone call you cannot see

The call metric on a profile counts taps on the call button, overwhelmingly from mobile. It misses the desktop user who copies the number, the person who writes it down and rings later, and every repeat customer with the number saved. It also counts wrong numbers, suppliers and hang-ups. Four practical ways to narrow the uncertainty:

  • A tracking number as the primary, with the real number kept as an additional number - the supported route to call data at profile level. It needs care, because the number on the profile should still match what appears on the website and citations, so this is a decision for the client rather than something you do quietly.
  • Sample rather than instrument - if permanent call tracking is not affordable, log the source of every call for two weeks each quarter. A fortnight of honest logging beats a year of guessing.
  • One question at the point of contact - "how did you find us" recorded in the booking system. Imperfect, cheap, and it captures the walk-ins and the word-of-mouth that no digital metric will ever show you.
  • Correlate rather than claim - line up call volume against the weeks you published posts, gained reviews or added services. If calls climb in the fortnight after review velocity doubles, say the two moved together and stop short of saying one caused the other.

Presenting the number to a sceptical client

Scepticism is reasonable. Most business owners have been shown a confident ROI figure by someone before you, and it did not survive contact with their bank account. Five things that hold up in the room:

  • Agree the model before the results exist - assumptions negotiated at proposal stage are shared. Assumptions introduced alongside a good month look invented.
  • Show the arithmetic - put the workings on the page. A model a client can audit is a model they can trust, even when the output is modest.
  • Give a range - "somewhere between 60 and 100 enquiries" is more credible than 102, and it is more honest about what the discount rates really are.
  • Point at the weaknesses first - name the double-counting risk and the customers who would have arrived anyway before the client does. It costs you nothing and buys the rest of the report.
  • Report cost per enquiry alongside their other channels - the figure means little in isolation and a great deal next to what they pay per lead on paid search or directories.

One structural note: build the report so the model is visible every month, not only at renewal. A client who has watched the same calculation run for six months argues about the inputs. A client shown it for the first time at renewal argues about whether to renew.

Frequently Asked Questions

Use a conservative flat discount on raw call volume, state it clearly in the report, and run a two-week manual log once a quarter to check whether the discount is roughly right. Adjust it at the review, not in the month it would flatter you.
As a diagnostic, yes. If enquiries fall while impressions hold steady, the problem is on the profile or in the competitive set rather than in visibility, and that distinction changes what you do next. It just should not be the number the client remembers.
Concede part of it immediately, because part of it is true. Then point at the segments where it is not: discovery searches rather than branded ones, and locations or services that were producing nothing before the work started. The strongest evidence is usually a new location or a newly listed service with a call volume that started at zero.
There is no universal figure, and anyone quoting one is guessing at your client's margins. The only comparison that means anything is against the same client's other acquisition channels and against their average job value. A cost per enquiry that looks high in one trade is excellent in another.
A5
Ampli5 Pulse Editorial Team
GBP Specialists · Ampli5 Pulse (Google Partner) · Ahmedabad, India
Our team has managed 4,000+ Google Business Profiles across 10+ countries since 2018. Every article comes from hands-on experience.
Ready to grow?

Manage Your GBP on Autopilot

AI review replies · Post scheduling · Performance insights · White-label reports. Start free.

Start Free Trial →