Every "best tools for your startup" list has the same problem: it is written from a company that already billed its first million. It recommends a CRM with automation rules to somebody with eleven customers and a spreadsheet that works perfectly well.
This is about the real first year. The one before the first hire, when every subscription comes out of the same account as payroll and every new tool is an afternoon not spent on product.
The rule before the list
A tool is worth it when it meets one of these three:
- It replaces work you are already doing by hand and that repeats every week.
- It does something you cannot do without it, not something you can do worse.
- Somebody requires it: the law, a customer, a bank.
Anything meeting none of the three is a bet on the future. Bets are allowed, but it helps to know that is what they are.
The second rule is less comfortable: the cost of a tool is not its price. It is the price plus the hours to set it up, plus the hours to maintain it, plus the cost of getting your data out the day you want to leave. A €29-a-month tool that takes two days to configure has cost, in month one, considerably more than €29.
The nine
1. A repository and its continuous integration
Not optional and not a decision. GitHub or GitLab, whichever you already know, with the actions included in the free plan.
What it costs: nothing at first. Free plans comfortably cover a team of one to five.
The usual mistake: building an elaborate deployment system before there is anybody to deploy to. A deploy that runs with one command in four minutes lasts far longer than it seems it should.
2. An infrastructure provider
AWS, Google Cloud, Hetzner, Railway, Fly. The choice matters less than the amount of arguing it attracts; what matters is not building something only its author can maintain.
What it costs: between €20 and €200 a month in the first year, depending on whether you need a managed database.
The usual mistake: the credits. Nearly every startup programme hands out €5,000 to €100,000 in credits, and that pushes people to build on proprietary services that get extremely expensive when the credit runs out. The credit is real; so is the bill afterwards.
3. Payments
Stripe, or Stripe. In Europe, with the tax situation of a SaaS selling across countries, there are not many serious alternatives that do not involve building a layer you do not want to build.
What it costs: a per-transaction fee. No fixed cost, so it does not appear until there is revenue, which is exactly right.
What gets underestimated: subscription state has to have one single source of truth. The moment your product keeps its own copy of "this customer is up to date" and somebody can edit it by hand, you start getting strange incidents nobody can reproduce.
4. Transactional email
The emails the product sends: verification, password reset, alerts. Resend, Postmark, SES.
What it costs: €0 to €20 a month at first-year volume.
The usual mistake: sending them from the same domain and the same reputation as the marketing emails. A campaign that gets marked as spam takes the password reset emails down with it, and that is a product incident.
5. Product analytics
Knowing what people do inside. PostHog, Amplitude, Mixpanel; or, at low volume, queries against your own database.
What it costs: the free tiers hold up well. PostHog's hosted free tier is generous.
What you actually need: three well-chosen events, not fifty badly placed ones. The event marking a customer's first result is worth more than all the rest combined, because it is the one that predicts renewal.
6. Somewhere to talk to customers
This is where I am writing from inside the industry, so take it with that caveat: what you need in year one is for conversations not to live in anybody's personal inbox. No more than that.
A shared inbox with a status and an owner covers 90% of what is needed until there are three of you. What does not work is the email alias: it distributes copies, not responsibility, and in the first week there is already a question nobody answered because each person assumed the other had it.
What it costs: €0 to €50 a month.
When you need more: when half the replies require data sitting on another screen. At that point the bottleneck stops being organisation.
7. Public documentation
An indexable help centre. It does not need an expensive tool: twenty well-written articles anywhere public beat two hundred on a platform with its own search.
Why it belongs in year one: because it is the only part of support that works while you sleep, and because it is indexable content. A share of any SaaS's search traffic ends up landing on its help articles.
8. Team password management
1Password, Bitwarden. Boring and not optional.
What it costs: €3 to €8 per person per month.
Why it cannot be skipped: because the alternative is a shared document, and that document is the first thing an audit asks for and the first thing to leak. With two people it already pays for itself.
9. Accounting
An accountant or software, depending on the country and the volume. In Spain, with electronic invoicing and the obligations coming into force, it is not territory for improvising.
What it costs: €60 to €250 a month.
What gets underestimated: the time. Handling invoices manually is three to six hours a month that disappear without anybody logging them anywhere.
The three you do not need yet
A CRM
With fewer than fifty customers, a spreadsheet with name, contact, signup date and status does exactly the same job, is faster to read, and needs no configuration.
A CRM starts earning its place when two people need to see the same state of the same opportunity and one of them was not on the call. Before that it is a place to put data that only its author reads.
Marketing automation
Sequences, lead scoring, branching flows. All of it solves a volume problem you do not have.
What does work in year one: writing emails by hand to the people who sign up. It does not scale, and that is why it works: it is the only way to find out why people arrive and why they leave.
A project management platform
With three people, a shared list and one conversation a day cover coordination entirely. A board with columns, tags, estimates and sprints is process for a problem that appears somewhere past eight or ten people.
The signal that you need one is not "we feel disorganised", it is "somebody did something somebody else had already done".
The question before signing
One, and it works for the nine and for the three:
Which specific task that I do this week disappears if I buy this?
If the answer takes a while to arrive, or if it is "we would be better organised", the tool does not replace work: it adds it. And in the first year of a SaaS the scarce resource is not money, it is the attention of the two or three people building it.
What is worth checking about everything you do buy, from day one, is how the data comes out. Not because you are planning to leave, but because knowing changes what you are willing to put in.
