Custom CRM vs Salesforce or HubSpot: A Decision Framework with Real Numbers

Choosing between a custom CRM and Salesforce or HubSpot comes down to three questions: how closely your process matches what a standard CRM models, what each option costs over three to five years (not just year one), and how much of the CRM’s value depends on data and logic that live in your other systems. If a standard CRM covers most of how you sell and you have fewer than a few dozen users, buy it. If the CRM is really part of your operations or product, or seat costs will dominate for years, building starts to make sense.

Most teams also have a third option they underrate: keep Salesforce or HubSpot and build the integrations that remove the manual work. This guide gives you a framework for all three, with current list prices, a cost formula you can fill in, and three examples from OSKI’s own projects.

Custom CRM vs Salesforce vs HubSpot at a glance

Current list prices (October 2026)

Salesforce Sales Cloud, list price as of October 2026, per user per month, as shown on Salesforce’s pricing page:

The page also lists a Free Suite at $0. Salesforce notes that AI can be added to Core and above, and that the page is for information only and subject to change. Its comparison table also lists some capabilities, such as API access on certain tiers and extra sandboxes, as paid add-ons, so price the edition you would actually need for integrations, not the entry tier.

HubSpot Sales Hub, as shown on HubSpot’s pricing page in October 2026:

• Free: free for up to 2 users.
• Starter: a low per-seat price; the page shows different figures by billing option, so check the current offer.
• Professional: starts at $90 per seat per month, with $100 shown for a different billing option, plus a required one-time Professional Onboarding fee of $1,500.
• Enterprise: starts at $150 per seat per month, plus a required one-time Enterprise Onboarding fee of $3,500.

HubSpot does not clearly label which Professional figure is the annual-commitment rate, so we use the lower “starts at” figure below and treat it as a floor.

Step 1: Process fit

Before any spreadsheet, map how your team actually works. Write down:

1. Core objects. Do you sell to accounts with contacts and opportunities, or do you manage something else: inspections, shipments, time logs, enrollments, projects?
2. The workflow that matters most. What happens from first contact to cash? Where do handoffs happen?
3. Rules that are hard to express as configuration. Pricing logic, approval chains, scheduling, pay or billing calculations.
4. Who uses it. Sales reps only, or also operations, finance, field staff and external partners?

Score the result honestly:

• Over 80% fit with standard objects: buy. Customizing a little is cheaper than building a lot.
• 50 to 80% fit: buy and integrate, or consider a platform build on Salesforce. Price the admin and developer time.
• Under 50% fit, or the “CRM” is really an operations system: a custom build deserves a serious estimate.

Worked example: OSKI Time & CRM

OSKI Time & CRM is a good example of a CRM where standard objects would not fit. Service companies bill clients and pay people by the hour, so every logged minute affects an invoice or a payslip. OSKI built its own platform that joins a desktop time tracker with a web CRM for worklogs, leave, Kanban boards, reports and wage or project pay calculation. The web CRM uses .NET, Angular and SQL Server, and the desktop tracker uses .NET MAUI with Blazor. One feature shows why custom made sense: before a log is uploaded, the tracker checks it against entries already in the CRM and, if times overlap, lets the user delete either the local log or the CRM entry, so each minute is counted once. OSKI runs its own team on it daily, and VMK Group UA has used it since 2023 for team time, leave and pay. In the words of VMK Group’s CEO on Clutch: “OSKI’s team presented us with a professionally crafted browser-based platform that they had developed and successfully utilized internally.”

Step 2: Total cost over 3 to 5 years

Comparing a year-one license bill with a build quote is the most common mistake in this decision. Use the same horizon for both.

The formula

SaaS CRM total cost

TCO_saas = Σ over years ( seats_y × price_per_seat × 12 )
         + one-time onboarding and implementation
         + Σ over years ( admin and configuration effort )
         + Σ over years ( add-ons and integration tooling )
         + integration build and upkeep

Custom CRM total cost

TCO_custom = discovery and build
           + Σ over years ( maintenance and change requests )
           + Σ over years ( hosting, monitoring, backups )
           + Σ over years ( support and security updates )
           + integration build and upkeep

Two terms deserve attention. Seats grow with the business, so use your headcount plan, not today’s number. And a custom CRM’s run cost scales with how often you change it, not with how many people log in.

An illustrative calculation

The following is illustrative only. The license prices are the October 2026 list prices above; every other figure is an assumption chosen to show the method, not a quote or a benchmark. Plug in your own numbers.

Assumptions: 40 users for five years, no headcount growth, no price increases, no discounts.

Now change one assumption: drop to 10 users. Salesforce Core licenses fall to $117,000 and HubSpot Professional to $54,000 over five years, while the custom build barely moves. At that size the custom option loses clearly, unless process fit forces it.

What the exercise shows:

• Seat count is the lever. Per-seat pricing is excellent value for small teams and becomes the largest cost for large ones.
• Admin effort does not disappear with SaaS. Someone configures fields, flows, permissions and reports. Budget it either way.
• Horizon matters. Over one or two years, buying almost always wins. Over five, the curves can cross.
• Risk is not in the table. A build can run late. A SaaS vendor can raise prices or retire a tier. Weigh both.

Step 3: Data ownership and hosting

With Salesforce or HubSpot, your data lives in the vendor’s cloud. You own it contractually, but access depends on their export tools, API limits and the edition you pay for. That is fine for most companies.

It matters more when:

• Clients or regulators require data to stay in a specific country, cloud account or on-premises network.
• The CRM holds data that your own product or analytics must query constantly, and API limits or extra fees get in the way.
• You plan to sell or white-label the system, which you cannot do with a SaaS CRM.

A custom CRM puts code and database under your control and lets you choose the hosting. The trade-off is that security, backups and uptime are now your responsibility, or your vendor’s under a support contract.

Step 4: Integrations

List every system the CRM must talk to: website forms, email marketing, accounting, ERP, support desk, telephony, your own product. Then for each, ask whether a ready connector exists for the SaaS CRM and what tier it requires.

Salesforce and HubSpot both have large marketplaces, which is a real advantage. But connectors cover common fields and flows. Anything specific to your business, such as custom pricing data, operational records or a homegrown back office, still needs integration work. Budget it in both columns of the cost table.

Step 5: Adoption risk

A CRM nobody uses costs the full price and returns nothing. Adoption risk looks different for each option:

• SaaS: familiar interfaces and plenty of training material, but reps may resist fields and steps that do not match how they work.
• Custom: screens can match the workflow exactly, but there is no outside training material, and a slow or buggy first release can lose users quickly.

Ways to reduce it: involve two or three daily users in design reviews, release a narrow first version that replaces one painful spreadsheet, and measure weekly active users from day one.

The hybrid option: keep the SaaS CRM, build the integrations

Many teams that think they need a custom CRM actually need their existing CRM connected to everything else. The hybrid approach keeps Salesforce or HubSpot for pipeline, forecasting and reporting, and adds custom integrations around it.

Worked example: Fairway Divorce

Fairway Divorce runs its website on Umbraco and .NET, and its sales and marketing teams work in Salesforce and Mailchimp. The problem was not the CRM; it was getting form submissions into it without retyping. OSKI wrote custom Umbraco Forms workflows that post submissions to Salesforce for sales and to Mailchimp for email marketing, with each form field mapped to the matching Salesforce and Mailchimp fields. The workflows sit inside Umbraco Forms, so existing form logic stayed as it was. Leads now reach the right teams with no manual data handling, and nobody had to replace a CRM.

The hybrid approach works best when:

• The CRM’s standard objects fit the sales process well.
• The pain is manual data movement between systems.
• Only some staff need CRM seats; others can work in the systems that feed it.

Our integrations service covers this pattern: forms, marketing tools, accounting and ERP connections around a CRM you already pay for.

When a custom CRM wins

Build when two or more of these are true:

1. The process is the product. Your workflow is how you compete, and standard objects force workarounds.
2. The CRM shares logic with operations. Time, pay, inventory, inspections or delivery data must live in the same system as customer data.
3. Seats dominate the long-run cost. Many users, a multi-year horizon and a stable process.
4. Data control is a requirement. Hosting, residency or resale rules rule out a SaaS vendor.
5. You have a clear specification and someone to own it. A product owner who can make decisions and review the work.

Worked example: an in-house CRM for Lokad

Lokad, a Paris-based company, chose to build its own CRM. According to Lokad CEO Joannes Vermorel’s Clutch review, OSKI built the in-house CRM from Lokad’s functional and architectural specifications between May 2023 and December 2024, with a team of 2 to 5 engineers, in a project Clutch sizes at $50K to $199K. Code was delivered through pull requests that Lokad peer-reviewed, following Lokad’s coding and architecture guidelines, and it was delivered on time. His summary: “They delivered a polished CRM, working fine and nice looking.”

Two lessons from that setup apply to any custom CRM: the client came with written specifications, and the client reviewed every change. That is what keeps a build on scope. You can read more about this kind of engagement on our CRM development page.

A decision checklist

Use this list with your team before you ask anyone for a quote:

• Process mapped from first contact to cash, with the hard rules written down.
• Fit score against standard CRM objects: over 80%, 50 to 80%, or under 50%.
• Seat count today and in three and five years.
• Five-year cost for each option using the formula above, with assumptions written next to every number.
• Integration list, with who builds and maintains each connection.
• Data hosting and residency requirements confirmed.
• Named product owner who will make decisions and accept releases.
• A first release scoped to replace one painful tool, not everything at once.

If the checklist points to buying, buy and invest in integrations. If it points to building, start with discovery and a narrow first release. Either way, the expensive mistake is deciding on year-one price alone.