Salesforce Consulting in Australia: Market Reality in 2026

Two winters ago I got called into a general insurer in North Sydney to look at a Sales Cloud build that a tier-one integrator had delivered over 14 months. The invoice total sat north of $2.3 million. The org had 340-odd active automations, four different trigger patterns fighting each other, and a Service Console that took nine seconds to load a single case. The head of technology asked me one question, and he asked it flatly, without heat: "Did we get ripped off, or is this just what Salesforce costs?"

Some version of that question is the most common thing Australian buyers put to me now. So here is my honest read on where the market sits in 2026, written from inside it rather than from a vendor deck.

The Market Split Into Two Halves And Hollowed Out The Middle

The Australian Salesforce market used to have a comfortable middle — mid-sized partners, 40 to 150 people, who could staff a genuine programme and still put an architect in the room who knew your org. That middle is mostly gone. It got acquired, or it got out-priced by the offshore-heavy model, or the good people left to go independent.

What's left is two halves. At the top you have the tier-one integrators and the global SIs, who win the enterprise and government panels because they can absorb the risk and tick the procurement boxes. At the other end you have boutiques and independents — often ex-tier-one architects who got tired of selling utilisation. There isn't much in between anymore, and buyers feel that as a cliff. You either pay for the brand and the bench, or you find someone who has actually shipped what you need.

I've watched this happen over the fifteen years I've been on the platform, and I wrote a longer piece on how that arc changed me in 15 Years on Salesforce. The short version: the platform got easier to build on and harder to build on well, and the market split along exactly that line.

Day Rates Say More About Risk Than Skill

Here's what rates actually look like in Sydney and Melbourne in 2026, from what I see quoted and what I quote myself.

An independent senior architect runs $1,400 to $1,900 a day. A boutique blended rate — architect plus a developer or two — lands somewhere around $1,600 to $2,100 a day depending on the mix. A tier-one blended rate on an enterprise programme is north of $2,400 a day, and that number carries a lot of people you will never meet.

The thing buyers get wrong is reading the day rate as a proxy for skill. It isn't. It's a proxy for risk transfer. When a bank pays a tier-one $2,400 a day, a chunk of that is insurance — someone to sue, a delivery methodology that satisfies the audit committee, and a bench deep enough that a resignation doesn't stall the programme. That's a real thing to buy, and for some organisations it's the right thing to buy.

But it is not the same as buying good engineering. On that insurance build, the $2.3 million bought me nine-second case loads and four trigger frameworks. I've since rebuilt the automation layer around a single pattern — the same one I describe in A Trigger Framework That Actually Scales — and the console now loads in under two seconds. The rate was never the problem. The seniority of the people actually touching the metadata was.

The Offshore Model Hit Its Ceiling

For a decade the dominant delivery model in this market was a thin local layer — a solution architect and an engagement manager onshore — sitting on top of a large offshore build team. It worked well enough for commodity Sales Cloud rollouts and it kept blended rates down. In 2026 it's running out of road, and not for the reason people usually cite.

The problem was never the location. It was the literalism. An offshore team paid on throughput builds exactly what the spec says, and Salesforce specs are always slightly wrong. You get code like this, which I have now seen in some form in more orgs than I'd like to admit:

// after-update handler on Case — one of four on the object
for (Case c : Trigger.new) {
    List<Contact> primary = [
        SELECT Id, Email FROM Contact
        WHERE AccountId = :c.AccountId AND Is_Primary__c = true
    ];
    // route Cases for a specific team to a queue
    if (UserInfo.getProfileId() == '00e5j000000AbCdEAAV') {
        c.OwnerId = resolveQueue(c);
    }
}

Three things are wrong in six lines: a SOQL query inside a loop that will die the moment someone bulk-updates cases, a hardcoded profile ID that means nothing to anyone six months later, and profile-name branching that should have been a custom permission. None of it is incompetence exactly. It's what you get when the person writing it was never in the room where the requirement was actually discussed, and had no mandate to push back. The fix for the query is a pattern anyone can learn; the deeper fix is judgement, applied at the point of writing.

That judgement is exactly the thing the offshore-throughput model priced out, and it's exactly the thing that's now getting expensive again.

Compliance Is The Real Moat In Australia

If you want to understand why the local boutiques still win work against cheaper delivery, look at compliance. This is where the Australian context stops being a footnote and becomes the whole game.

APRA CPS 234 makes the regulated entity — the bank, the insurer, the super fund — accountable for the information security of anything a third party touches. Data residency assumptions get written into contracts. Government work runs through IRAP assessments and, increasingly, questions about where every byte of PII physically sits and which support engineer can see it. None of that is hard to say. It's very hard to actually deliver against while also shipping a working org, and it does not survive being handed to a team that has never read the Australian Privacy Principles.

I've had two engagements in the last three years where the entire vendor selection came down to a single question in the security workshop: can you demonstrate that no production data leaves Australian-controlled infrastructure during support. The cheaper bid couldn't answer it cleanly, and the price difference stopped mattering. That's the moat. It isn't glamorous, and it's the reason a Sydney boutique can quote above an offshore blended rate and still win.

AI Changed What You're Actually Paying For

The commodity end of this work has collapsed in price, and it collapsed fast. The Apex boilerplate, the test class scaffolding, the first-draft LWC, the migration mapping — the things that used to fill an offshore timesheet — are now something I generate in an afternoon with the right tooling. I wrote about what that does to the whole business model in How AI Changes the Salesforce Consulting Model.

The consequence for buyers is specific. You should no longer be paying senior rates for volume. If a proposal is priced on how many developer-days it takes to produce a known quantity of code, you are buying a thing that got cheap. What hasn't got cheap — what actually got more valuable — is the judgement about what to build, what to refuse to build, and how to keep the org maintainable after the consultants leave. Pay for that. Stop paying for keystrokes.

I'll state the opinion plainly, because it's the one I argue most often. In 2026 the right engagement is smaller and more senior than the one you were sold in 2020. Fewer people, more experience, shorter timeline, and a deliberate handover to your own team. Anyone quoting you a large blended team for a mid-sized build is quoting you the old model, and the old model is subsidising juniors on your budget.

When A Boutique Is The Wrong Call

I run a boutique practice, so take this with the appropriate salt — but a boutique is genuinely the wrong choice for some work, and pretending otherwise is how buyers get burned.

If you're running a multi-year, multi-cloud transformation across a workforce of thousands, you need bench depth that an independent or a five-person shop simply cannot provide. A key-person resignation shouldn't be able to stall a $20 million programme, and with a small team it can. If your procurement function mandates a certain revenue threshold, insurance level, or panel membership, that's a hard constraint and no amount of engineering quality gets you around it. And if the work is genuinely about change management across a large, resistant organisation — not the Salesforce build, the humans — then the delivery methodology and the sheer number of hands a tier-one brings are worth what they cost.

The mistake isn't hiring a tier-one. The mistake is hiring one for a job that needed three senior people for four months, and then wondering where $2.3 million went. If you want a second pair of eyes on which of those two situations you're actually in before you sign, that's the kind of review I do.

The North Sydney insurer didn't get ripped off, exactly, and they didn't just pay what Salesforce costs. They bought the wrong shape of team for the problem they had. Two years on, their console loads in under two seconds and a team of their own people maintains it — which is the only outcome that was ever worth the money.