A few years back, I found myself knee-deep in a CRM nightmare, juggling multiple systems that just wouldn't play nice. We tried HubSpot for a client project, it was good, but not quite. Then, Salesforce for another, powerful, yes, but overkill for what we needed. I finally got around to convincing my team to build a custom CRM. The decision boiled down to pure maths: costs vs. adaptability. Using Airtable and Zapier, we tracked every penny and every hour saved. Now, in 2026, the numbers speak volumes.
The Actual Numbers Nobody Posts About
Let's start with seat costs, because that's where most people focus. Fair enough. But seat costs alone will mislead you badly.
HubSpot's Sales Hub Professional runs around $90 per seat per month in 2026 (they've nudged prices up again). Five seats. That's $450/month or $5,400/year before you've touched a single add-on. Add the Marketing Hub if you want email sequences tied to your pipeline and you're looking at another $800/month minimum. Salesforce Sales Cloud Professional sits at roughly $80 per seat, but nobody actually uses Salesforce at Professional tier without buying CPQ, Einstein Analytics, or a third-party Salesforce implementation partner. That implementation alone cost one of our fintech clients at Seahawk somewhere north of £18,000 before they'd sent a single email through it.
Pipedrive is the honest one of the three. Around $49 per seat at the Advanced tier, clean UI, no nonsense. For a five-person sales team that just needs pipeline visibility, it's genuinely hard to argue against.
Here's the thing though. None of these sticker prices include:
- Data migration. Moving contacts, deals, and notes from wherever you are now is always a project. Budget 20-40 hours if you're doing it properly.
- Training time. Salesforce has a learning curve that would humble a PhD student. HubSpot is friendlier but still takes two to three weeks before anyone's using it correctly.
- Workflow lock-in. Your processes change. Every change in HubSpot or Salesforce means either a developer, a paid consultant, or three hours of your own afternoon gone.
Year-One vs Year-Three Cost is What Matters
Year one, HubSpot looks tolerable. Year three, after two price hikes and four modules you added because the sales rep convinced you they were "essential for scale," the number looks completely different.
I tracked this for a twelve-person agency we supported through a CRM migration in 2024. Their HubSpot bill had gone from £680/month in year one to £1,940/month by month thirty-one. No new seats. Just feature creep and a price increase they agreed to in a renewal email nobody read carefully.
---
What HubSpot Is Actually Good At
Genuinely. I'm not here to trash it.
HubSpot wins when your marketing team and your sales team need to share the same platform without a developer gluing things together. The contact timeline, the email open tracking, the sequence builder, when it works, it works beautifully. And the free CRM tier is legitimately useful for solo operators or tiny teams who just need a place to log calls and track deal stages.
The HubSpot Academy is also genuinely one of the better free educational resources in the industry. I've pointed junior account managers there dozens of times.
Where HubSpot falls apart: complex B2B sales with non-standard pipelines, anything that requires deep custom objects without a developer, and any business where the pricing model means you're paying for contacts you'll never email. That contact-based pricing is a slow burn. You hit 50,000 contacts and suddenly your monthly bill has a number in it that makes you sit down.
---
The Salesforce Problem (And Who It's Actually For)
Salesforce is extraordinary software. It can do almost anything. That's also the problem.
Back in 2021, a logistics client came to us after six months of a failed Salesforce implementation. A big consultancy had scoped it, sold them on the full platform, and then delivered something so customised and so brittle that when one of the consultancy's developers left, nobody could maintain it. They were paying £3,200/month in licences for a system their team had stopped using by month four.
Salesforce makes sense at scale. We're talking 50+ seat sales orgs, enterprise-grade compliance requirements, teams that have a dedicated Salesforce admin (a real one, not "Steve from IT who did the Trailhead modules"). If you're below that threshold, you're paying for runway you'll never use.
One genuine strength: the Salesforce AppExchange ecosystem is enormous. If your industry has a niche workflow, someone's probably built an extension for it. That has real value.
---
Pipedrive: The Underrated Option
I'll be honest, I underestimated Pipedrive for years. Wrote it off as "HubSpot lite." That was wrong of me.
For sales-first teams where the pipeline is the product, Pipedrive's visual deal board is excellent. The automation at Advanced tier handles 80% of what most small-to-mid agencies actually need. And the price-to-value at five to ten seats is genuinely difficult to beat.
Seahawk uses a hybrid setup internally: Pipedrive for outbound sales tracking, connected to our project management layer via Make (formerly Integromat). Took a developer one day to set up. It's been running clean for eighteen months.
Where Pipedrive loses: marketing automation is thin, reporting is limited unless you export to a BI tool, and customer support quality has dropped noticeably since their 2023 growth push. If you need deep segmentation or behaviour-based email triggers, you'll be duct-taping a second tool on top.
---
The Case for Building Custom in 2026
This is where most articles wave their hands and say "it depends." I'll be more specific.
Custom makes sense when three things are true simultaneously:
- Your workflow genuinely doesn't fit inside a standard pipeline/contact/deal model
- You have at least 10 seats and a three-year planning horizon
- You can allocate £15,000-£40,000 upfront, or have in-house development capacity
That last one kills most of the conversation for smaller operators. And fairly so. But if those three things are true, the maths over five years is often decisive.
A custom CRM built on something like Directus as a headless data layer, with a React or Next.js front-end, and Resend for transactional emails, runs maybe £25,000-£35,000 to build properly. No per-seat fees. No price hikes. You own the data model completely.
We built exactly this for a property management client in Bristol. Forty-two users, very specific tenancy workflow that no off-the-shelf tool handled cleanly. Two years post-launch, they're saving approximately £4,100/month compared to their old Salesforce stack. The break-even hit at month eleven.
What Custom Actually Costs to Maintain
This is the part people skip. A custom CRM is not free after build. You need:
- A developer on retainer or in-house (budget £1,500-£3,000/month for occasional updates and security patches)
- Hosting (Render, Railway, or a managed VPS: £80-£200/month depending on traffic)
- Monitoring and backup tooling (another £50-£100/month if you're doing it properly)
Add that up. Around £2,000/month ongoing at the low end. For ten seats that's £200/seat, more expensive than Pipedrive, cheaper than HubSpot with add-ons, and dramatically cheaper than Salesforce at equivalent feature depth.
---
The Decision Framework I Actually Use
When a client asks me which direction to go, I run through this in roughly this order:
- How many seats, now and in eighteen months? Below eight seats and you're not building custom unless you have unusual requirements.
- Is marketing automation core to the CRM, or handled elsewhere? If core: HubSpot deserves serious consideration. If elsewhere: Pipedrive or custom.
- What does your pipeline actually look like? One linear pipeline with five stages? Any tool works. Multi-branch, role-dependent, with conditional automations? You're hitting the limits of off-the-shelf fast.
- What's your tolerance for vendor dependency? Some founders hate the idea of a price hike decision sitting with a San Francisco product team. Those founders should build custom.
- Do you have a developer, or access to one? No developer, no real custom option. Full stop.
The answer for most ten-to-twenty person agencies in 2026 is Pipedrive (connect it to Make for automations) or a lightly customised HubSpot free tier plus one paid module. Not Salesforce. Not custom, unless the workflow genuinely demands it.
---
FAQ
Is HubSpot free tier actually usable in 2026?
For solo operators or very small teams tracking under 1,000 contacts, yes. It covers basic deal stages, contact logging, and email integration. The moment you need sequences, reporting beyond surface level, or more than one pipeline, you'll hit the paid wall fast. Budget for at least the Starter tier if you're using it as a real sales tool.
How long does a custom CRM take to build?
Properly? Twelve to twenty weeks for a first version that your team can actually use in production. I've seen agencies promise eight weeks and deliver something broken at week ten. If someone quotes you six weeks for a full custom CRM with automations and reporting, ask hard questions about scope.
Can Pipedrive replace Salesforce for a 30-person sales team?
Depends entirely on what that team actually uses Salesforce for. If they're using basic pipeline management and email logging, yes, Pipedrive handles that and saves them a significant amount monthly. If they're using CPQ, custom approval workflows, or deep ERP integration, Pipedrive won't cover it without significant additional tooling.
What's the biggest hidden cost across all of these?
Time. Not the software cost. The hours your team spends working around a tool that doesn't quite fit. I've seen salespeople at clients spending forty minutes a day on manual data entry because the CRM automation wasn't set up right. That's over three hours a week per person. At ten people, that's thirty hours a week of productivity quietly leaking out of the business. No line item for it on the invoice.
Ultimately, whether you lean towards a custom CRM or one of the big names, it's all about fit. I've learned that sometimes the 'biggest' isn't the 'best',it's the solution that works seamlessly with your unique workflow. So before diving in, crunch your own numbers. Maybe even grab a coffee first. You'll need it.