How to Build an Apollo Sequence With a Manual Research Step

How do you build an Apollo sequence that starts with real research?

Make the first step an action item that asks you to research the account, then follow it with a manual email that uses what you found. Only after that personal first touch should automatic emails take over for follow-ups. The research step slows the start slightly but makes every later email land better.

This tutorial is for one situation: you are a founder or a solo seller running outbound to a small list of target accounts each week, maybe twenty to fifty. You use Apollo for sequences. Your problem is that fully automatic sequences feel generic, and fully manual outreach takes too long to keep up.

I build outbound systems for B2B teams, and this hybrid pattern is the one I recommend most often for small teams. It keeps the human judgment where it matters, at the first touch, and lets automation handle the reminders and follow-ups that do not need it.

What step types does Apollo give you to work with?

Apollo describes four core step types in its own guide: automatic email, manual email, phone call, and action item. Automatic emails send based on parameters you set. Manual emails create a scheduled task to edit and send a customized message. Phone call steps hold notes and scripts. Action items are fully customizable tasks.

The action item is the key to this setup. Apollo describes it as a customizable task for internal steps along the way to a sale. That makes it a natural place to put "research this account before writing."

Apollo's interface and options change over time, so check Apollo's own documentation for the exact menus and any newer step types before you build. The structure below works with the four step types described above.

What should the research step ask you to do?

Give yourself a short, fixed checklist inside the action item: read the company's homepage and one recent post or update, check the contact's role and how long they have held it, and write one sentence on why your offer might matter to them now. Cap it at ten minutes per account.

The cap matters. Without it, research becomes procrastination, and the sequence stalls. Ten minutes is enough to find one real reason to write. If you cannot find one in that time, that is useful information too. The account may not be a fit right now.

Write your one sentence in a note on the contact, so the manual email step can use it. That note also helps later, when the prospect replies and you need to remember why you reached out.

A good research sentence is concrete and checkable. Something like "They just opened a second sales role, so lead routing is probably getting messy" is useful. "They seem like an innovative company" is not. If the sentence would not change what you write in the email, keep looking or move on to the next account.

How should the manual first email be written?

Short, specific, and built around the one sentence from your research. Open with the reason you are writing to this person now, connect it to a problem you solve, and ask one easy question. Keep it under a hundred words. Use a template for structure, but rewrite the opening every time.

A good structure is: observation, relevance, question. The observation comes from your research. The relevance is one line about how you help companies in that situation. The question is low commitment, like whether the problem is on their radar this quarter.

Avoid pasting in claims you cannot back up. If your research found nothing specific, it is better to write a short, honest, role-based email than to invent a compliment. I made the same point in what to check before an AI SDR sends for you.

When should automatic emails take over?

From the second email onward. Once the first touch is personal, follow-ups can be templated, because their job is to bring the first message back to the top of the inbox and add one new piece of value. Two or three automatic follow-ups spaced several days apart is plenty for a small list.

Each follow-up should add something: a short example, a relevant resource, or a different angle on the same problem. "Just bumping this" adds nothing and trains the reader to ignore you.

Set the sequence to stop when someone replies. Check Apollo's documentation for how reply handling works in your account, and test it with a colleague before you load real prospects.

Where do phone calls fit in?

Add a phone call step after the first or second email only for your best-fit accounts. Use the step's notes to paste your research sentence and a two-line opener. A call that references the email you sent feels like a follow-up, not a cold interruption.

Many small teams skip calls entirely, and that is a reasonable choice. If you do call, keep the list short and the bar high. Five well-prepared calls a week will teach you more than fifty rushed ones.

If phone is not part of your motion, replace this step with a second action item to check whether the account has done anything new since your first email.

How do you get the right contacts into the sequence?

Start with a short, verified list of the right person at each target account. Bad contacts waste the research step, because you spend ten minutes on someone who left the company or never owned the problem. Build the list carefully before you build the sequence.

If you use Clay for list building, my tutorial on building a Clay table that finds the right contact covers how to get one or two verified buyers per account. The same principles apply to any data source: tight titles, low limits, and a manual review.

Also make sure your sending domain is ready before the first email goes out. My guide to cold email domain setup covers the basics.

How do you know if the sequence is working?

Track positive replies per hundred contacts and meetings booked, not open rates. Compare accounts where your research found a strong reason against those where it did not. If the strong-reason accounts reply much more often, your research step is earning its time. If not, tighten what you look for.

Review the sequence every two weeks. Read the replies you got and the ones you did not. Most improvements come from changing the first email, because that is where the research shows up.

Keep the research notes. Over time they become a picture of which reasons to write actually lead to conversations, and that is the best input for your next sequence.

What should you do next?

Build a four-step sequence in Apollo: an action item for ten minutes of research, a manual first email built on that research, and two automatic follow-ups that each add something new. Load ten accounts, run it for two weeks, and read every reply before you scale it up.

If you want help designing an outbound sequence that feels personal without eating your whole week, reach out. I am happy to look at your list and your first email.