Cold email software: how to pick your stack

Most people shopping for cold email software start with a feature list. A better starting point is the work itself: who finds the contacts, who decides they are a fit, who writes, what sends, and who answers the replies. Every tool covers some of those jobs and leaves the rest to you.

This guide walks through the jobs, the four kinds of tools on the market, what the big mailbox providers require before you send a single message, and three stacks that fit different team sizes. It names real tools where their own websites describe them, and it leaves prices out because they change too often to be useful here.

In short:

• Map the five jobs first. Buy tools for the jobs nobody on your team wants to do by hand.
• Deliverability and the law are part of the stack. Gmail’s rules and the consent rules in your recipients’ country decide what you can send.
• Fewer tools means fewer gaps. Every hand-off between two tools is a place where data goes stale.

The five jobs of an outbound stack

Whatever you buy, someone or something has to do these five jobs. Writing them down shows you where your gaps are before you compare cold email tools.

1. Find the companies and the right person inside each one. You need a name, a role and a way to reach them.
2. Decide who is a fit. Not everyone in a search result is a good prospect. This is where an ideal client profile earns its keep.
3. Write a message about the prospect’s situation. One relevant observation beats three paragraphs of features.
4. Send safely from a mailbox that is set up, warmed up and watched, with an unsubscribe link and a list of people who asked not to be contacted.
5. Handle replies. Sort the interested from the “not now”, answer quickly, and keep a record of who said what.

Under all five sits a sixth job that is easy to miss: keeping one record of every company and person, so you never write to the same prospect twice from two tools.

The four kinds of cold email software

Tools on the market fall into four rough groups. Many tools sit across two groups, so read the table as a rough map.

Data and enrichment tools

These tools start from a database or from several data providers. Apollo describes itself as one connected go-to-market system, with a B2B contact database, outreach sequences, a dialer and AI agents. Clay combines multiple data providers for better coverage and lets you research target companies and people with AI.

A data tool is the right first purchase when finding the right person is your bottleneck. It is the wrong one when you already know your market well and the hard part is writing and following up.

Sending tools

Sending tools exist to get a sequence out of many mailboxes on a schedule and to keep those mailboxes healthy. Instantly presents itself as sales engagement and lead intelligence: it finds leads, builds AI sales agents and automates outreach.

Warm-up means sending a small, growing number of ordinary emails from a new mailbox and getting replies to them, so mailbox providers learn that the address belongs to a real sender. Most sending tools do this for you in the background.

A sending tool needs a steady supply of contacts and good copy. Without those, it only makes it faster to send messages nobody wanted.

Multichannel tools

Multichannel tools put several channels into one sequence. lemlist says it reaches leads via email, LinkedIn, calls, WhatsApp and SMS.

Each extra channel brings its own rules. LinkedIn, for example, limits invitations and forbids unauthorised automation in its User Agreement; our guide to safe LinkedIn limits sums up what LinkedIn says officially.

All-in-one platforms and AI agents

The newest group tries to cover all five jobs in one place, often with an AI agent that searches, writes and drafts replies. You get fewer tools to connect and fewer gaps between steps, and you give up some freedom to swap one part for a specialist.

If you go this way, check who approves what. An agent that sends replies on its own is a different product from one that drafts and waits for your click.

What Gmail requires before you send

The mailbox providers decide whether your email lands, so their rules belong in the stack decision. Google publishes sender guidelines for messages going to personal Gmail accounts. According to Google’s email sender guidelines, every sender must:

• set up SPF or DKIM for the sending domain,
• keep the spam rate reported in Postmaster Tools below 0.3%,
• send over a TLS connection,
• have valid forward and reverse DNS records.

Senders of more than 5,000 messages a day to Gmail must also set up DMARC, use both SPF and DKIM, align the From domain with one of them, and support one-click unsubscribe for marketing messages.

You may never reach 5,000 messages a day. Meet the stricter list anyway. It costs little, and every serious inbox provider is moving toward that standard.

What this means for the tool choice:

• The tool must send from a mailbox and domain you control, so the authentication records are yours.
• It must add an unsubscribe option to every message and remember who used it.
• It should warn you before you send to a list that is likely to bounce or complain, because a high spam rate damages your domain’s reputation, and that stays with you when you change tools.

Questions to ask before you buy a cold email tool

A demo shows the tool on a good day. These questions cover what happens on the other days.

1. Whose mailbox does it send from? Your own mailbox and domain keep your reputation in your hands. A shared sending pool mixes your reputation with strangers’.
2. How does it handle unsubscribes and suppression? Ask what happens when the same person appears in a new list next month.
3. Who approves replies? Find out whether replies can go out without a person, and whether that is on by default.
4. Does it write well in your prospects’ language? A message with a wrong form of address or a broken name reads as spam in any language.
5. Where is the data hosted? If you sell in the EU, ask where contact data and message history live.
6. How is it priced? Per seat, per mailbox, per contact or with credits. Work out what your normal month would use, and check what sending costs on its own.
7. Can you leave? Check that you can export contacts, companies and history as CSV.
8. What happens when something goes wrong? Ask how the tool tells you about bounces, a blocked mailbox or a sudden rise in complaints, and whether it pauses sending on its own.
9. Does it work with your AI assistant? Some tools now offer an MCP server, so you can run them from Claude or ChatGPT. See prospecting with Claude for what that does and does not do.

Three stacks by team size

Pick the row closest to your situation and adjust it.

A simple test helps when you are torn between two options: count the hand-offs. Every time data moves from one tool to the next, someone has to check it moved correctly. Three hand-offs a week is manageable. Ten is a part-time job.

How to test a tool before you commit

A trial week tells you little if you test the wrong thing. Most tools look good on a fresh account with ten hand-picked contacts. Test the boring parts instead.

1. Use a real segment. Take one slice of your market that you know well, such as one industry in one country, and run it through the whole flow. You will spot wrong contacts faster in a market you know.
2. Read every message before it goes out. Count how many you would send without editing. That number tells you more about the writing quality than any demo.
3. Check the data where it is weakest. Look at small companies and at countries the vendor is not based in. Many databases are good for large firms in large markets and thin everywhere else.
4. Unsubscribe from your own test campaign. Then add the same address to a new list and see whether the tool stops you.
5. Reply to your own test email from a second mailbox, once as an interested buyer and once with “not now”. See where each reply lands and what the tool suggests.
6. Export everything at the end. If the export is messy now, it will be worse after a year of data.
7. Ask support one real question during the trial, such as how to move your unsubscribe list in from another tool. How fast and how clearly they answer is a fair preview of the help you will get once you pay. Ask too which languages and hours support covers.

Write down what you find for each point. When you compare two tools later, you will compare notes instead of memories of two different demos.

Common mistakes small teams make with the stack

Buying data before you know who you want. A database with millions of rows does not help until you can describe your ideal client in a few lines. Write the profile first, then buy the data that covers it.

Sending from your main domain on day one. A new campaign with a list you have not checked can push your bounce and complaint rates up. Warm up a mailbox, start with a small daily volume, and watch the numbers for a couple of weeks.

Treating every reply the same. “Not interested”, “talk to my colleague” and “send me details” need different answers. If your tool cannot tell them apart, set a fixed time each day to sort replies by hand.

Keeping two lists of who opted out. When the sending tool and the CRM each keep their own list, someone will eventually be written to after they asked to stop. Pick one place and make every tool read from it.

Measuring opens instead of conversations. Open tracking is unreliable because many mail apps load images automatically or block them. Count replies, positive replies and meetings instead.

The law shapes the stack too

The sender decides who gets a message, and the rules for that differ by country. No tool takes that decision off your hands.

Under the GDPR, Recital 47 says processing personal data for direct marketing may be regarded as a legitimate interest, and Article 21 gives every person the right to object to direct marketing at any time, after which their data may no longer be used for it. That is why a working suppression list is not optional.

Email rules sit on top of the GDPR and vary by country. In Czechia (Act No. 480/2004 Coll.) and Germany (UWG, section 7), for example, commercial email may go only to people who gave prior consent or to an existing customer you write to about your own similar products or services. This applies to business addresses too. Read our guide to cold email law in B2B before you choose who goes into a campaign.

For the stack, this means two features you cannot skip: an unsubscribe link in every message, and one suppression list that every campaign and every tool respects.

HypeLead in your outbound stack

HypeLead belongs to the all-in-one group. It searches 400M+ contacts in our global search and Czech company registers by IČO, judges each lead against your ideal client profile, and looks for buying signals such as hiring, funding, web and news. It writes personal messages in natural Czech or English, with formal address and correctly declined names in Czech, and sends from your own mailbox and your own LinkedIn. Before anything goes out it checks SPF, DKIM and DMARC, warms the mailbox in a warm-up pool and checks the spam score. Every email carries a one-click unsubscribe. Replies land in one inbox, drafted by AI and, by default, sent on your click.

You can see the whole flow on the product page and how credits work on the pricing page. If you are comparing it with a specific tool, the comparison pages list what each one does better.

---

Want one tool for search, writing, safe sending and replies? Your next step is at the end of this page.