A Marketing Stack You Can Actually Read

Key takeaways

• Think in layers, not logos: record, measure, reach, enrich, bill.
• One source of truth per layer. Two CRMs is worse than one mediocre one.
• Most early-stage needs are covered by free tiers; the cost of a stack is mostly the time it takes to keep it clean.
• Add a tool when a question you ask every week can't be answered, not when a competitor uses it.
• The tools record what happened. Something still has to connect those records to what you did and tell you what to do next.

View full size

Ask a five-person SaaS team how many marketing tools they pay for and the answer is usually a pause, then a number higher than the team size. Each tool made sense when it arrived. Together they produce a dashboard per tool and no shared answer to the question that matters on Monday morning, which is what to do this week.

The five layers

Every B2B SaaS stack, however large, reduces to the same five jobs.

If a tool doesn't fit one of these jobs, ask what question it answers before you buy it.

Stage 0: before product-market fit (1 to 3 people)

You need very little, and most of it is free.

• Record: a CRM on a free tier. HubSpot, Pipedrive and Attio all work at this size; pick the one your sales process fits and stop comparing.
• Measure: GA4 and Google Search Console, both free. Set up conversion events for the two or three things that matter, such as signups and demo requests, and leave the rest alone.
• Reach: whatever channel you are testing, usually founder-led outbound, a few posts a week and maybe a small paid test.
• Bill: your payment provider's own dashboard.
• Decide: a spreadsheet with one line per marketing action. The action log guide shows the columns.

The main risk at this stage is buying tools to feel organised. Ten conversations with buyers will teach you more than any of them.

Stage 1: early traction (3 to 10 people)

Now you have repeatable channels and a funnel worth measuring. Add tools in this order, and only as the need shows up.

1. Email and automation. A newsletter and simple nurture flows. Most CRMs include basic email; use that first.
2. Ads accounts with conversion tracking. Connect the conversions back to your analytics so you're not judging campaigns on clicks.
3. Enrichment. Once inbound volume makes manual research slow, an enrichment tool fills in company size and role so you can tell good leads from noise.
4. Product analytics, if you're product-led. When the signup-to-active step matters, you need events inside the product, not just page views.
5. A CMS your marketers can edit without a developer, if the website is still in code.

Stage 2: scaling what works

At this point the questions get harder: which channel brings customers who stay, which content brings pipeline, what a lead is worth.

• Self-reported attribution. A required "How did you hear about us?" field on signup or demo forms. It's cheap and often catches what tracking misses, like podcasts, word of mouth and AI assistants. The direct traffic guide explains why tracking misses so much.
• A reporting layer. When people start exporting CSVs from three tools to answer one question, a BI tool or Looker Studio (free) saves real time.
• Revenue in the CRM. Pull billing data into the CRM so deals show actual revenue and churn, not just "closed won".

What to skip early

Some categories are useful at scale and expensive distractions before it.

• Multi-touch attribution suites. With a few dozen conversions a month, the models don't have enough data to say anything stable.
• Intent data. Useful when you have a sales team ready to act on it, and noise before.
• Customer data platforms (CDPs). They solve the problem of many data sources feeding many tools. With five tools, direct integrations are simpler.
• A second tool for the same layer. Two analytics tools that disagree will cost you more time than either saves.

How to connect the layers

The value of a stack is in the joins, not the tools. Four habits keep the joins working.

One source of truth per layer. Decide which tool owns contacts, which owns revenue and which owns web behaviour. When numbers disagree, you know which one wins.

A UTM naming rule, written down. Lowercase, one spelling per source, a campaign name that matches the action. Without it, "LinkedIn", "linkedin" and "li" become three channels in every report.

Measure across domains. If signup or checkout happens on another domain, set up cross-domain measurement in GA4 so the source survives the hop.

Send the same IDs everywhere. The CRM contact, the analytics user and the billing customer should be linkable, usually through the email at signup or a user ID, so a deal can be traced back to where it started.

View full size

The decision layer

Each layer answers its own question. The CRM says how many deals moved, analytics says how many people came, billing says who paid. None of them says whether the webinar last month worked, or which of this quarter's five actions to keep. That needs a place where the funnel numbers from all five layers sit next to a list of what the team did, compared with what's usual for each stage.

There are three common ways to build it.

• A spreadsheet. Weekly numbers per stage pasted in, an action log next to them. Free and flexible, and it works as long as someone keeps it up.
• A BI tool. Looker Studio, Metabase or similar, connected to your sources. Better charts and less manual work, but you design the comparisons yourself, and the action log usually lives somewhere else.
• Sporely. This is the tool we build. It connects to the CRM, analytics and billing layers, keeps the action log, and checks each action against the usual range of the stage it was meant to move. It's one option among the three; for a team with very few actions a spreadsheet may be all you need.

Whichever you choose, the test is the same: on Monday morning, can you see which stage is leaking and which of last month's actions moved anything?

Common traps

• Buying for the team you'll be. An enterprise stack for a three-person team means most of it goes unused and none of it is set up well.
• Integrations nobody owns. A sync that breaks silently is worse than no sync. Give each one an owner.
• Measuring everything. Twenty conversion events make the five that matter harder to find.
• Letting each tool define the funnel. If the CRM, analytics and ad platforms each count "leads" differently, agree on one definition per stage and map each tool to it.

Quick Q&A

What marketing tools does a B2B SaaS startup need?

At the start, a CRM, GA4 and Search Console, the channels you are testing and your billing provider's dashboard. Most have free tiers. Add email automation, enrichment and product analytics as traction grows.

How many marketing tools should a small team use?

One per layer is a good rule: record, measure, reach, enrich and bill. For a team under ten, that's usually five to eight tools including ad accounts.

Do I need a marketing attribution tool?

Not early. With few conversions, multi-touch models aren't stable. A "How did you hear about us?" field and consistent UTMs answer most questions until volume grows.

Should marketing and sales share a CRM?

Yes. One CRM as the source of truth for contacts and deals keeps lead counts consistent and lets you trace revenue back to where a lead started.

How do I connect my CRM, analytics and billing data?

Use the same identifier everywhere, usually the email or a user ID set at signup, keep UTM naming consistent, and decide which tool owns each number when they disagree.

Keep reading

• How to decide what marketing to do next: channels, fair test windows and when to look at competitors.
• Track marketing actions, not just traffic: a simple action log, reviewed weekly.
• Why your converters show up as direct traffic: where the source gets lost.