Salesforce

Where Salesforce sits in the market

Salesforce has been the reference point for CRM for so long that reviewing it feels less like reviewing a product and more like reviewing an institution. It tops most CRM market-share tables, it sells to everyone from two-person startups to the largest enterprises on the planet, and its pricing page is a moving target of editions and add-ons that I have learned to read like a contract rather than a menu. When I look at a CRM at this size, the question is never "is Salesforce good?" — of course the functionality is there. The question is whether the platform commitment is worth what it costs a given team in licences, implementation, and the admin hours nobody puts in the business case.

The 2026 story at Salesforce is Agentforce, the company's agentic AI layer, and the pitch has moved decisively from "system of record" to "platform where AI agents do work alongside your team." Whether that pitch lands for your business depends almost entirely on how much of your process already lives inside Salesforce — AI features are multipliers on an existing data discipline, not substitutes for one. The company is also in the middle of an edition renaming exercise (Core, Advanced and Max now occupy the ground Enterprise and Unlimited used to hold), which means a lot of the written material you will find online — including pricing comparisons from even two years ago — describes tiers that no longer exist under those names.

What you actually get

Sales Cloud is sold in editions, and the ladder has been re-cut recently. At time of writing, salesforce.com/sales/pricing lists a Free Suite for up to 2 users, then Starter Suite at $25 per user per month, Pro Suite at $100, Core at $195, Advanced at $395, and Max at $550, all per user per month billed annually. Those numbers are list prices; negotiated enterprise deals deviate from them in both directions, and Salesforce's own sales team will happily tell you the list price is "a starting point." In practice I expect mid-market buyers to land meaningfully below list on Core and above list on the add-ons.

Functionally, the core is what it has been for years: a relational data model built around accounts, contacts, leads and opportunities, with campaigns, quotes, forecasting, and territory management layered on depending on edition. On top of that sit the parts that actually differentiate Salesforce in 2026. Flow, the automation builder, is genuinely capable — record-triggered flows, scheduled paths, screen flows and now Agentforce actions cover a lot of what used to require Apex code. The AppExchange ecosystem is the largest in the category by a wide margin, which matters more than any feature list: the odd vertical requirement you have has probably been packaged and sold by someone already. Einstein and Agentforce bring predictive scoring, generative drafting, and now autonomous "agents" that can handle service cases or sales follow-up actions within guardrails. Data Cloud exists to make all of that AI work against unified data across systems, and it is priced separately on consumption, which is the single most important thing to understand about Salesforce's current commercial model: the seat price is no longer the whole bill.

Underneath the marketing, the platform fundamentals are strong. The API layer is mature and well documented, identity and permission management is granular to the point of being a profession in itself, and the platform handles multi-currency, multi-language and complex sharing rules without blinking. If your requirement is "model an unusual sales process exactly, and change it every year," Salesforce is one of maybe three products in the world I would trust to do that for a decade.

My testing notes

In my testing of a fresh Sales Cloud org over a couple of weeks, several things stood out.

First, the good: Flow has quietly become the centre of gravity. I set up lead routing — assign inbound leads round-robin to a small team, escalate anything untouched after four hours, and post a Slack alert on escalation. The round-robin plus escalation flow took me a focused afternoon, all in the declarative builder, and the Slack step was two more clicks because of the native integration. Two years ago that recipe reliably pushed you into Apex code and a developer; now it does not. The step from "toy automation" to "production automation" is still steep — testing, deployment between environments, and permission handling all require the discipline Salesforce's own admin certifications pretend you have — but the ceiling of what a capable admin can do without code is much higher than the category average.

Second, the uncomfortable: the edition gaps are where the product's personality shows. On Starter Suite I could build pipelines and basic reports and felt deliberately fenced in — validation rules, deeper automation, and sandbox-style experimentation live up the ladder. The practical consequence is that "Salesforce at $25 per user" and "Salesforce the platform" are two different products, and teams routinely discover this mid-implementation, after the migration. I also tested Einstein-generated opportunity summaries on a populated org: the output is decent when your data is clean and close to useless when it is not, which is a fair summary of Salesforce AI generally. If your reps have been treating required fields as suggestions, the AI will faithfully summarise the mess.

Third, from my notes: the report builder is powerful once the data model is right and punishing before it is. I built the same pipeline-ageing report twice — once against a tidy org with consistent stages, once against an imported spreadsheet dump — and got a useful dashboard and a confusing one from identical clicks. The tool was not the variable.

Fourth, and this shapes how I advise buyers: the AppExchange is a genuine strength with a hidden cost. I installed two well-known add-ons during testing, and both multiplied the configuration surface — new objects, new permissions, new places where things can silently break during upgrades. The ecosystem extends what Salesforce can do; it also extends what your admin has to understand.

Living with it: contracts, partners, and the hidden curriculum

Salesforce buying has its own craft. Multi-year commitments trade price for lock-in; fiscal-quarter-end timing genuinely affects what sales will agree to; and add-on SKUs are where budgets leak. I have seen teams negotiate hard on seats and then accept list on Data Cloud consumption eighteen months later. If you buy Salesforce, get the renewal uplift caps in writing at the first signature, not at renewal.

Implementation is the other half of the cost. A competent partner implementation for a mid-market sales team is commonly a five-figure project before anyone logs in, and the alternative — self-implementation — is viable only if someone inside the business can commit serious time to learning the platform's conventions. The training ecosystem (Trailhead, certifications, community) is the best in enterprise software, which is its own kind of admission that you will need it.

Limitations

Cost at scale is the obvious one, and it is not just the seat price. Add-ons — Sales Engagement, CPQ (now Revenue Cloud), Data Cloud consumption, Agentforce conversation credits — are separate SKUs, and a serious Salesforce org typically carries several of them. A team of 50 on Core at list is roughly $117,000 a year before a single add-on or implementation hour; that math has to pay for itself in pipeline, and plenty of mid-market teams find it does not.

Complexity and admin overhead are the second. Salesforce is an ecosystem you staff, not an app you install. Small teams without a dedicated admin should think hard before committing, because unmaintained Salesforce orgs accumulate technical debt faster than almost any other software I have worked with — dead fields, contradictory validation rules, flows nobody remembers writing.

Third, the reporting and analytics experience still carries legacy weight. The classic report builder feels dated, and the newer analytics products are, predictably, paid upgrades. API limits and storage limits also bite larger orgs at surprising moments, usually during a data migration. And while reliability at the platform level is strong (the company publishes trust data and takes it seriously), individual org slowness is frequently self-inflicted through customisation, which is a problem the buyer owns.

Finally, the AI story ahead of the fine print. Agentforce is real technology and a real commercial push, but the practical deployments I consider successful are narrow: well-scoped service deflection, clean opportunity summaries, guided follow-up. Treat the broader "agentic enterprise" framing as roadmap, not as something your org gets on day one.

Verdict

Salesforce earns its position when you need a platform: a data model that will survive five years of process change, an integration ecosystem that ends most "can it connect to X" debates, and AI features that compound over data you have actually cleaned up. It is the wrong answer for a ten-person sales team that wants a good CRM this quarter — the cost, the admin burden, and the gap between cheap editions and the real product all point elsewhere (Pipedrive, Close, and HubSpot's lower tiers are more honest fits there).

My rating balances what the platform can do against what it costs and demands. Functionally it is near the ceiling of the category; on value for money and usability for the uninitiated, it pays a tax. At time of writing, list pricing runs $25 to $550 per user per month for Sales Cloud depending on edition, and the deals actually signed rarely match the sticker either way.