Setting up a ticketing system at an SME sounds easy until you sit down to compare the options. Zendesk, Freshdesk, HubSpot Service, Zoho Desk, Jira Service Management, Front, Help Scout… then there are open-source projects like osTicket, GLPI or Zammad, which look free but cost you time. And in the background, always, the question of whether you should build the whole thing yourself.
Some teams pay for more help desk features than they use, while others handle every issue in a shared spreadsheet. The decision should start from your ticket volume, your processes and the integrations you need.
To choose a customer service system, you need to compare processes, integrations and how you expect the team to evolve. This guide brings together the criteria for assessing a SaaS product, an open-source option or a custom build.
What a ticketing system really does (and what it doesn't)
A ticketing system is, essentially, a shared inbox on steroids. Every customer request (an email, a WhatsApp message, a web form, a phone call) becomes a ticket with an ID, a status, an owner and a history. So far, nothing anyone wouldn't understand.
What we tend to forget is that a modern ticketing system does a lot more: it routes requests automatically by rules (if the customer is premium, it goes to team A), applies different SLAs depending on the contract, keeps auditable records, connects to the CRM and ERP so the agent can see who's on the other end, and lets you measure the team's real productivity. It's the difference between firefighting and running customer service as a measurable process.
What it WON'T do for you
This is where many SMEs get a nasty surprise. Customer service software doesn't fix broken processes: if your team doesn't know who answers what, giving them Zendesk won't solve it, it'll just digitize it. Nor does it replace a good knowledge base: if your agents keep procedures in their heads, you'll have the same problems, just with prettier dashboards.
A help desk is only as good as the processes it automates. If your processes are chaos, the software amplifies the chaos.
The market options: SaaS, open source and custom
Let's cut to the chase. Today, the serious options for a Spanish SME fall into three families.
Commercial SaaS
Zendesk, Freshdesk, Zoho Desk, HubSpot Service Hub, Intercom, Front and Help Scout offer different plans by agent and feature set. Check what each one includes and which add-ons you need for chat, AI or telephony.
Pros: you're up and running in days, there's support in Spanish (not always, so watch out with Zendesk on some plans) and constant updates. Cons: you pay per agent forever, and prices go up every year.
Self-hosted open source
osTicket, GLPI, Zammad, FreeScout and UVdesk are open-source options. You need to plan for installation, configuration, updates and backups, handled by an in-house team or a maintenance provider.
Custom development
A custom help desk requires an upfront investment plus maintenance. Weigh its total cost against the licenses and integrations of the SaaS alternatives, taking into account code ownership and your team's needs.
When a standard SaaS is all you need
Let me be clear: most Spanish SMEs don't need a custom system. If your situation matches the following, sign up for a SaaS and stop overthinking it:
- A customer service team of 1 to 8 people.
- Standard processes (you receive requests, categorize them, answer them, close them).
- Steady volume, with no huge spikes.
- Integrations the SaaS covers out of the box: Outlook, Gmail, Slack, WhatsApp Business, Shopify, WooCommerce.
- You don't handle particularly sensitive data or have unusual industry requirements.
In that scenario, a SaaS such as Freshdesk or Zoho Desk can meet your needs without custom development. Check the integrations, and invest time in training the team and documenting procedures.
The classic mistake when choosing a SaaS
Note that the big mistake isn't choosing the "wrong" SaaS, it's signing up for the top-tier plan before you know which features you'll use. Start with the mid-tier plan, give it three months, and upgrade only if the data justifies the premium features. Interestingly, 60% of companies pay for features they never get around to switching on.
Signs you've outgrown your help desk
Now for the interesting part. There comes a point when a SaaS, however good, stops working for your business. It isn't the tool's fault: your operations have simply become unique. These are the signs we see again and again at Interconecta when a company calls us because "Zendesk isn't enough anymore":
- Licenses take up an ever-bigger share of the budget. Compare the total cost over the next few years with that of your own solution, including its maintenance.
- You have unusual approval flows that don't fit the SaaS rules without wrestling with APIs and Zapier.
- You need deep integrations with your ERP (Sage 50, Holded, Odoo, SAP B1) that the standard connector doesn't cover.
- Sensitive data or a regulated industry: healthcare, banking, legal. The GDPR and Spain's Data Protection Act (LOPDGDD) require you to have full control over where the data lives, and with a US-based SaaS that's always an issue.
- Your product is complex and the agent needs to see, inside the ticket, customer data, invoice history, contract status, telemetry from the installed equipment, and so on.
- You switch SaaS providers every two years because none of them quite fits. That's the clearest sign of all.
If you've ticked three or more, you're no longer in "just get a SaaS" territory. You're looking at a custom help desk or, at the very least, your own software layer on top of the SaaS.
What a well-built custom ticketing system includes
Custom development isn't "let's copy Zendesk, but uglier". That's what happens to people who hire on price alone. A custom ticketing system makes sense when it really takes advantage of being yours:
Native integration with your stack
The agent opens the ticket and sees, without extra clicks, the customer's last three invoices, the status of their active contracts, open issues in other departments and the sales rep's upcoming visits. With a SaaS you can approximate this with widgets, but it's never quite the same.
Real business rules
If your company has different SLAs by customer type, time of day, product purchased, contract signed… a custom build encodes all that without contortions. In Zendesk, you end up with 47 triggers and a separate spreadsheet to make sense of them.
AI tailored to your use case
This is where it gets interesting. Your own help desk lets you plug in AI models (automatic ticket classification, suggested replies based on your documentation, urgency detection through tone analysis) tuned to your industry's vocabulary. If this approach interests you, take a look at how we use AI chatbots for business connected to internal systems.
Serious compliance
Servers in Spain or the EU, encryption at rest, detailed audit logs, integration with Veri*Factu (Spain's new invoicing compliance system) when a ticket generates an invoice, and data retention configurable by customer type. Things that, with a standard SaaS, mean endless negotiations.
How much a custom help desk costs and how long it takes
The cost depends on the scope. These three levels help define what the proposal should include:
Working MVP
A basic ticketing system with authentication, a multichannel inbox (email + web form), manual and automatic assignment using simple rules, statuses, SLAs, a dashboard and basic metrics. 2 to 3 months of development. Suitable for teams of up to 10 agents with processes that aren't too complex.
Advanced help desk
Everything above, plus: integration with your existing ERP/CRM, a mobile app for field technicians, a customer portal, a knowledge base, CSAT/NPS surveys, complex automations and a billable-hours module. 4 to 7 months. Designed for professional services firms, SaaS technical support or industrial maintenance.
Custom AI platform
When you add automatic classification with your own models, suggested replies, IP telephony integration, call transcription and analysis, and a real-time executive dashboard. From this point on, every project is different.
Include corrective and evolutionary maintenance in the proposal. It should cover incidents, updates and integrations, with clear responsibilities and terms.
Custom development isn't an expense, it's an asset. But only if it's maintained. Without maintenance, in five years you'll have thrown your money away.
The middle ground: SaaS + your own layer
There's an option many people don't consider, and in my experience it's the one that works best for companies in transition. Instead of choosing between SaaS and custom, you combine both: you keep the SaaS for agents' day-to-day work (because its interface is well designed and your people already know it) and build your own layer on top that connects to your ERP, automates complex rules and adds tailored dashboards.
A maintenance company can keep its help desk and add its own layer to turn alarms into tickets, assign technicians and generate signed work reports. That layer is priced according to the integrations and workflows it needs to cover.
This combination fits especially well with customer service automation, where you don't need to reinvent the help desk but do need to integrate it with the rest of your operations. It's usually quicker to roll out and less risky than a full build from scratch.
How to decide without getting it wrong: the honest checklist
Before you sign anything, SaaS or custom, answer these questions in writing. If you don't have clear answers, you're not ready to choose a tool: you're ready to review your processes first.
- How many tickets do you handle per month? How many do you expect in 24 months?
- How many people handle them? Will that number grow?
- Which channels do you really need (not the ones that would be nice to have)?
- Which SLAs have you committed to in your customer contracts?
- Which internal systems must the help desk talk to?
- What budget do you have for the next five years, not just the first?
- Who will maintain this when the person who chose it leaves?
If, after answering all of this, you realize your case is standard, go with a SaaS. If you find three or four specific needs that no product covers well, consider your own layer or a full custom build. And if you're unsure, ask two SaaS vendors and one custom development partner for fixed-price quotes: compare like with like over five years, not just the first year.
One last piece of practical advice
Don't start with the software. Start by sketching on a sheet of paper the five most common types of request you receive, who handles them today, how long it takes and what steps are involved. If you can't sketch that in an hour, any tool you buy will be an expensive patch. If you can and it's clear, you already know what your ticketing system has to do: run that workflow reliably, measurably and at scale.
Choosing a ticketing system isn't about deciding between Zendesk and Freshdesk. It's about deciding how you want your company to serve its customers over the next five years. If your volume is low and your processes are standard, a well-configured SaaS will serve you without drama. If your operations are unique, or the license fees are starting to hurt, consider something custom.
What we usually see at Interconecta is that most SMEs start with a SaaS, stick with it for two or three years, and then rethink. That isn't a failure of the SaaS: it's proof that the company has matured and needs a tool that matches. If that's where you are, read our articles on customer service automation and custom web applications before you sign another three-year renewal.
And if you're unsure, send us an email, tell us about your case and we'll tell you honestly whether you need custom development or just need to fine-tune what you already have. Sometimes the most honest answer is: stay where you are.