Pricing a GoHighLevel SaaS mode plan that survives usage

SaaS mode is the point where an agency stops billing for its time and starts billing for access. That is a genuinely better business — right up until the first month where the usage charges arrive and the margin you thought you had turns out to belong to a telecom carrier.

The fix is not a higher price. It is a plan structure that separates the three things you are actually selling.

The three things inside every SaaS plan

1. Access — a sub-account on the platform, under your brand. This costs you a share of a fixed monthly platform fee, and the cost per client falls every time you add a client.
2. Usage — calls, texts, emails, AI minutes. This is variable, it is entirely driven by how successful the client is, and it is billed to you before it is billed to them.
3. The build and the babysitting — the system inside the sub-account, plus whoever keeps it working. This is the part with real cost and the part clients actually value.

Price those three as one number and you have a plan that is unprofitable for your busiest clients and overpriced for your quietest. Price them separately and the plan self-corrects.

Start with the platform floor

The platform tier that unlocks SaaS mode, connected billing and automated provisioning is published by GoHighLevel and is a fixed monthly cost to your agency, not a per-client one. That single fact is the entire reason the model works: your access cost per client is that fixed number divided by however many clients you have.

At two clients that division is brutal. At twenty it is a rounding error. So the correct way to price the access component is not against today’s client count but against the count you are underwriting the plan to — and then to accept that the first handful of clients are subsidised by you rather than by the plan.

Then deal with usage honestly

Usage is where SaaS-mode plans quietly die. A client who runs a reactivation campaign across four thousand contacts consumes in a fortnight what your flat plan assumed they would consume in a year.

There are exactly three defensible answers, and mixing them is fine:

Whichever you choose, publish the multiple. A markup on a pass-through cost is defensible when it is stated and indefensible when it is discovered.

A worked scenario — substitute your own numbers

Assume you underwrite the plan at twelve clients. Assume the platform tier is a fixed monthly cost to you, so your access cost per client is that figure divided by twelve. Assume the average client sends a moderate volume of texts and takes a moderate volume of AI-answered calls, and that you rebill that usage at a stated multiple. Assume you spend, on average, forty minutes a month per account on maintenance and one hour a quarter on a review call.

Now write the four lines out: access cost, usage cost, your own labour at whatever you actually value an hour at, and the price. The gap is your margin. Two things fall out of that sum almost every time.

• The labour line, not the platform line, is the largest cost. Agencies obsess over the platform fee because it arrives as one visible invoice, while the forty minutes a month goes unbilled and unnoticed.
• Margin per client improves with every client added, but only until the labour line crosses your own capacity. Past that point each new client costs a fraction of a hire, and the curve inverts.

That inversion is the real reason agencies plateau in SaaS mode, and it is why the labour line is the one worth fixing first — either by making the build identical enough that maintenance is trivial, which is what a standardised white-label snapshot does, or by converting the hours into a fixed monthly line through white-label fulfilment.

Designing the three tiers

Three is the right number, and the middle one should be the one you want sold.

• Entry — access, the core build, a modest usage allowance. Its job is to make the middle tier look sensible, not to be profitable at volume.
• Standard — the full system, a realistic allowance, priority support, and whatever monthly touch you actually intend to deliver. This is the plan. Price it against the client’s outcome, not against your cost.
• Advanced — everything, plus the things that consume your time unpredictably: extra automations, integrations, campaign builds. Price it so you are indifferent to whether it sells.

Do not build a tier around a feature the client cannot feel. Tiers built on message limits get compared to telecom pricing. Tiers built on what the system does for the client’s customers get compared to nothing at all.

What the tiers should be described in terms of

Your client is not buying seats. They are buying the fact that a stranger who calls their business at nine at night gets a real conversation and a booked appointment rather than voicemail. They are buying reminders that go out the day before and an hour before, and an automatic call within the hour when somebody does not turn up — a no-show recovery process built to a seventy per cent recovery target. They are buying a review request after every completed job, aimed at a steady five to ten reviews a month.

Write your tiers in those terms and the price stops being compared against a software subscription. It gets compared against the cost of the missed calls, which is a number your client already knows and already resents.

Three ways a SaaS-mode plan leaks margin

Each of these is quiet, which is why they are worth naming. None of them shows up as a bad month; they show up as a business that is busier than last year and no more profitable.

• The unpriced favour. A client asks for one small automation and gets it free because it takes twenty minutes. Repeated across twenty clients across a year, that is a part-time job nobody is paying for. Have a named place for small requests — a monthly allowance, or a rate — so the answer is a process rather than a mood.
• The grandfathered plan. Early clients are on early pricing, and nobody raises it because they were loyal. Two years later they are the accounts consuming the most support at the lowest price. Raise annually, in writing, attached to something new.
• The allowance that was never revisited. A client whose volume tripled because the system worked is now consuming three times the usage on the same plan. Success is supposed to trigger a tier change; build the review into the calendar or it will not happen.

The common thread is that all three are consequences of the system working well. That is worth sitting with — the failure mode of a good deployment is an unprofitable one, unless the plan anticipates growth rather than assuming stability.

The short version

Cover access with volume, cover usage with a stated multiple, and cover labour by making the build the same every time. Price the middle tier against what the system does for your client’s customers, and let the other two exist to make it obvious.