Marketing Tech Stack: How to Build a B2B Martech Stack From the Data Up

What is a marketing tech stack?

A marketing tech stack is the set of software a marketing team uses to reach, convert and measure its audience, connected so that data flows from one tool to the next. It usually includes a CRM, a website and content system, email and automation, advertising, analytics and a data layer that says who your accounts and contacts are.

You will also see it called a martech stack or a marketing technology stack. The three terms mean the same thing. Martech is short for marketing technology; adtech is the narrower part of it used to buy and serve ads. A martech stack can contain adtech, but most of what a B2B team runs every day (CRM, email, content, analytics) is not advertising technology.

The word "stack" matters. A list of subscriptions is not a stack. Tools become a stack when they share the same records: the person who downloads a guide, opens an email, clicks an ad and books a demo should be one contact with one history, not four partial copies in four tools. Most of the work of building a stack is making that true.

This guide covers the layers of a B2B martech stack, why stacks fail, why the data layer should come first, three example stacks by team size, how to build one in six steps, how to audit the one you have, and where AI assistants now plug in.

The 6 layers of a B2B martech stack

Every B2B marketing tech stack, from a founder with three tools to a team of fifty with thirty, can be read as six layers. The table gives the role of each and the question it should answer.

Two layers carry the others. The data layer is what every other tool reads from and writes to: if a contact's company or title is wrong there, the email goes to the wrong segment, the ad audience is off and the dashboard counts the wrong accounts. The analytics layer is what tells you whether the rest is worth paying for. The four layers in the middle are where most of the budget goes, and where most stacks start.

A short note on integration tools. Connectors such as Zapier, Make or n8n are not a layer of their own; they are the wiring between layers. They matter a lot in small stacks, where they replace native integrations, and they become a source of quiet breakage in large ones.

The minimum martech stack for a B2B team

If you are starting from nothing, the minimum marketing technology stack for a B2B team is shorter than most vendor diagrams suggest. You need one system of record for accounts and contacts (a CRM, or a single spreadsheet at the very start), a way to complete those records, a website with forms that write to the record, one email tool, and web analytics. That is five pieces of martech, and a small team can run a full year of campaigns on them.

Everything else is added when a specific job appears: a scheduling tool when demos start, an SEO tool when content becomes a channel, an ad platform when there is budget to test, a sales engagement tool when outbound grows beyond what one person can send by hand. Adding martech before the job exists is the fastest way to end up with licences nobody opens.

Martech, adtech and salestech

The boundaries between categories are blurry, and they matter mostly for budgets and ownership. Martech covers the tools marketing owns; adtech the subset used to buy media; salestech the tools sales owns (pipeline, outreach, call recording). In B2B the three share the same records, which is why the data layer belongs to neither team alone: when marketing and sales each keep their own copy of an account, the marketing stack and the sales stack disagree from the first week.

Why most martech stacks fail

Stacks rarely fail because a tool is bad. They fail for three reasons that have little to do with any single tool.

1. The data is fragmented. Each tool keeps its own copy of the contact. The CRM has the title from two years ago, the email tool has the title from the signup form, the ad platform has none. Segments disagree, reports do not match, and nobody trusts the numbers. In our experience, fragmented data is the most common reason a stack disappoints, and adding tools makes it worse.
2. The tools are not connected. A tool that does not send its data back to the CRM is a silo. Webinar attendees who never reach the CRM cannot be followed up by sales; ad clicks that never reach analytics cannot be measured. A tool that does not connect does not count as part of the stack, whatever it costs.
3. Nobody uses them. Tools are bought for a project, the person who chose them leaves, and the licences keep renewing. Low adoption is invisible on the invoice and obvious in the audit (see the audit checklist further down).

All three point the same way. The order in which you build the stack matters more than the brand of each tool. A team that starts from the data layer, connects each new tool to it and checks adoption before renewing often gets more from five tools than another team gets from twenty.

Start with the data layer

The data layer answers three questions: where do your accounts and contacts come from, what do you know about each one, and how does that information stay current?

Where records come from

In a B2B stack, records arrive from many places: website forms, product signups, event lists, LinkedIn, purchased or built prospect lists, and imports from sales. Each source brings different fields. A form gives an email and a name; a signup may give only an email; a prospect list gives a company and a title but no history. Decide which tool is the system of record (usually the CRM) and make every source write to it.

What you know about each record

Most marketing decisions need a few fields that sources rarely provide: company, industry, headcount, country, the person's title and seniority, and sometimes the tools the company already uses. Segmentation, lead scoring, ad audiences and reporting by segment all depend on them. When these fields are empty, every layer above works on guesses.

This is where enrichment comes in: starting from what you have (a name and a company, a LinkedIn URL, a company domain) and adding the missing fields from public sources. Derrick is a data enrichment platform that does this in its web app, in Google Sheets, in Claude, ChatGPT or any MCP client, and through a REST API. In practice:

• Import a list of contacts into the web app and run Enrich Leads to add title, company and location from LinkedIn, at 1 credit per profile.
• Run Enrich Companies on your accounts to add industry, headcount, headquarters and website from each company's LinkedIn page, at 1 credit per company.
• Use Website Technologies (2 credits per website, paid plans) to see which tools an account's website runs, useful when your product integrates with a specific platform.
• Add work emails with Email Finder (5 credits per email found, paid plans) when a list needs to reach the engagement layer.

The LinkedIn-based features need your LinkedIn account connected through the Derrick Chrome extension. The free plan gives 100 credits a month, enough to test enrichment on a sample of your CRM before deciding anything.

How information stays current

People change jobs, companies grow, titles change. A record that was right a year ago is partly wrong today. Plan a refresh: re-enrich active records on a schedule (every quarter is a common rhythm), and re-check a record when it re-enters a campaign. Teams with a developer can call the API (Plus plan and above) to enrich new records as they arrive and refresh old ones in the background; others can export from the CRM, enrich in the web app or in Google Sheets, and re-import as an update.

For a deeper look at running marketing from a clean database, see our guide to B2B database marketing.

B2B marketing tech stack examples by team size

Whatever the size, the data row decides whether the others work. The three stacks below are illustrative examples, not surveys of real companies. They show how the same six layers fill up as a B2B team grows, using tool categories rather than brands so you can map them to what you already pay for.

The founder alone

Five or six tools, mostly free or cheap. The spreadsheet is the CRM, and that is fine as long as one sheet is the single source. The founder builds a list of target accounts, enriches it in the Derrick web app or in Google Sheets, and works it by hand. The mistake at this stage is buying a large automation platform before there are enough contacts to automate.

The team of 5

The first real CRM arrives, and with it the first integration work. Every form, signup and list now writes to the CRM. Enrichment becomes a habit on every new list and every new signup batch, so segments and lead scores have the fields they need. This is also when tools start to overlap: two email tools, two form builders. The quarterly audit begins here.

The team of 50

Thirty tools or more, several owners, and a dedicated person for marketing operations. Data no longer moves by hand: new records are enriched automatically through an API before they reach the CRM, and active records are refreshed on a schedule. The risk changes from missing tools to unused ones, and from empty fields to conflicting ones. Governance (who can buy a tool, who owns each integration) becomes part of the stack.

Whatever the size, the data row is the one that decides whether the others work. A team of five with a clean CRM will run better campaigns than a team of fifty whose tools disagree on who the customer is. For the capture and engagement layers in detail, see our guides to lead generation tools and sales automation tools.

How to build a martech stack in 6 steps

Build in this order: data first, then only the missing tools, then the connections, then measurement.

1. Write the goals and the motion. What should marketing produce this year (signups, demos, pipeline on target accounts), and how do customers buy? A product sold through demos to a few hundred accounts needs a different stack than a self-serve product with thousands of signups. If you sell software by subscription, our guide to B2B SaaS marketing shows how the acquisition model shapes the plan.
2. Audit what you already have. List every tool, its owner, its cost and who actually uses it. You may already own most of what you need (the audit method is below).
3. Fix the data layer. Choose the system of record, define the fields every record should have, and fill them for the accounts and contacts you already hold. Measure the fill rate of each key field before and after.
4. Add tools only where a layer is missing. For each gap, write the job it should do and the metric it should move. Buy for the job, not for the feature list.
5. Connect every tool to the system of record. Before you sign, check that the tool reads from and writes back to your CRM, natively or through a connector. If it cannot, treat it as a temporary experiment, not as part of the stack.
6. Measure, then review every quarter. Set up the analytics layer so each tool can be judged on the metric you wrote in step 4, and put the audit in the calendar.

The order is the point. Many teams start at step 4 and discover steps 3 and 5 a year later, when the reports do not match.

How to audit your martech stack, tool by tool

A martech audit is an inventory with a decision on each line. Strategy questions (who the stack serves, what it should achieve) come first, but the savings and the fixes come from going through the tools one by one. Build a sheet with one row per tool and these columns.

Then check the data layer itself. Export a sample of contacts and accounts from the CRM and measure, for each key field (company, industry, headcount, title, country), the share of records where it is filled and the share where it is still right. A quick way to check correctness is to enrich the sample again and compare: the fields that changed show how fast your data ages. Run the audit every quarter; the first one usually finds tools to cancel, the following ones mostly find data to fix.

Where AI agents plug into the stack

AI assistants are becoming a working surface of the marketing tech stack, not a separate tool. A marketer asks an assistant to build a segment, draft a sequence, summarize a campaign or research an account, and the assistant needs to read and write data to do it.

The Model Context Protocol (MCP) is the standard that lets an assistant such as Claude or ChatGPT call tools directly. Through Derrick's MCP server, a marketer can ask "who leads marketing at these five target accounts?" or "add headcount and industry to this list" in a conversation, and get enriched data back without leaving the chat. MCP and API access start on the Plus plan.

The rule from the rest of this guide still holds. An assistant is only as good as the records it reads. Connecting AI to a fragmented, outdated CRM produces confident answers about the wrong people. Fix the data layer first, then give assistants access to it.

Common marketing tech stack mistakes

• Buying the execution layer first. An automation platform on top of an empty CRM sends well-designed emails to badly segmented lists.
• Counting tools instead of jobs. The question is not how many tools you have but whether each job (capture, engage, publish, advertise, measure) has one clear owner and one tool.
• Accepting tools that do not connect. Every tool that keeps its data to itself creates a second version of the truth.
• Never refreshing the data. Enriching a list once and never again turns a clean database into a stale one within a year.
• Renewing by default. Without a quarterly audit, unused licences renew for years.
• Letting each team buy alone. Marketing, sales and customer success buying overlapping tools on their own budgets is how a team of fifty ends up with three email tools.

Your marketing tech stack on one page

If you keep one thing from this guide, keep the order: data first, then capture, engagement, content and advertising, then measurement, with every tool connected to the same records. Write your stack on one page, with a line per layer, the tool that fills it, its owner and the metric it moves, and review it every quarter.

For the full list of what Derrick can add to a record, see the features page. And if you want the wider picture of how marketing teams in this space use data, browse the B2B marketing guides.