Instagram DM Opener: Why the Standard Follow Message Sounds Like a Bot

The standard Instagram DM opener sounds like a bot because people have already received it dozens of times. “Hey, thanks for the follow! Are you more into X or Y?” was taught in so many business courses that a new follower recognizes the pattern from the preview line alone and moves on. With openers like that, many of our clients start at around 20% replies, sometimes 9 or 10%. A short, casual message that looks typed by a person in a coffee queue often gets twice as many.

Twice as many replies is twice as many conversations. With the same conversion from conversation to booked call, that is twice as many sales calls from the same followers. That is why I fix the opener first with every client, and treat it as a test that never ends.

Why the standard follow message gets ignored

It’s banner blindness. When you see the same billboard on your way to work every day, after a while you stop seeing it at all. When it changes, you notice it again.

Welcome messages work the same way. What a new follower sees in the inbox preview is “Hey, thanks for following” and nothing else. They have seen that line from many accounts over the years, so the brain files it as an automated message that doesn’t need a reply. Ads that open with “If you’re a coach, trainer or expert…” get the same reaction: boredom.

There’s nothing fundamentally wrong with the thank-you format. Some clients start at 35% with it because their audience still responds to it. But for most accounts it is the most worn-out pattern in the inbox, and the data says so too.

What the data says about openers

We looked at 292 welcome message versions from 91 Setor accounts, 723,857 sends in total. To avoid comparing a strong account with a weak one, the comparison only counts accounts that ran both kinds of opener.

Short wins, a question helps, thanking for the follow and mentioning it hurt. That matches what I see in client accounts. Messages under eight words, written in lowercase and loose, like you’re quickly greeting a new follower while waiting for your coffee, often do twice as well as a block of text that looks pasted.

For reference: on half of active accounts at least 26.9% of people reply to the welcome, 34.5% on the top quarter and 44.1% on the top tenth. More on how the welcome works is on the Instagram welcome message page.

What to send instead

When I started my business, before Setor existed, I messaged a lot of people by hand, and I wasn’t going to write essays to people who might never reply. So I sent one word: “you there?” A lot of people answered.

It works because it implies something is going on. When someone calls and asks “can you talk right now?”, you assume they have something to tell you. The opener uses the same small sense of urgency. On my own profile, “you there?” got 65% replies at first. It has since settled around 50%, which is normal: at the start you message people who followed minutes ago, and later you work down to people who followed a while back and are colder. The same goes for your script: a drop from 15% to 7.5% after a few weeks usually doesn’t mean something got worse, only that you’re now messaging colder people.

A few openers in the same spirit:

you there?

can I ask you something?

can I give you feedback on your profile?

can I be honest?

you know what’s interesting about your profile?

Please don’t copy these word for word. Other people read the same advice, some of them will copy them, and the pool of followers in popular niches will have seen them two or three times by next week. That is banner blindness again, just with a newer template.

Three rules I give clients when we write their first variants:

1. Start with a hook, not with your question. Most templates go straight to “are you more into X or Y?”. A hook only has to get a reaction. The questions, lead magnets and offer come later in the conversation.
2. Skip “hey”. A huge number of template messages start with it. I literally tell clients to start in the middle of the sentence: drop “hey, can I ask you something” to “can I ask you something”.
3. Write it the way you’d text a client. You don’t need to be a copywriting expert. My best ads were often the ones I wrote on the fly, at the gym or on a walk, because they sounded natural. If you’re struggling to come up with something, you’re probably trying too hard. Take a walk, think about your clients and write what you’d send to get a reaction.

Your clients aren’t stupid. When they get several paragraphs with capital letters, full punctuation and line breaks, they know you didn’t just write it because they became your 20,000th follower. It’s a template, and templates don’t need a reply.

Run an A/B test from day one

An A/B test means sending opener A to one follower and opener B to the next, alternating. After a few hundred sends you can see, for example, that 20% reply to A and 40% reply to B. Start with two variants. If you test three, split traffic roughly 33/33/34. In Setor you do this in the welcome message A/B tab, where a slider sets the share of people who get each variant.

The rule that saves the most money: don’t draw conclusions before each variant has about 200 replies. Replies, meaning started conversations, not sent messages. At a 40% reply rate that’s around 500 sends per variant.

With 10 conversations on each side, one early booking can make a variant look like the winner just because it converted first. Switch off the other one and you may have killed the variant that would have been twice as good over a month. If there are fewer than 200 replies, don’t touch the test, even if you’re tempted.

Judge by booked calls, not only replies

One of our clients in fitness, an account we had worked with for months, ran three openers. All three got about 37% replies: 36, 37 and 38%. But two of them booked about 1.5 qualified calls per 100 welcomes, and the third booked 3. Same reply rate, twice the bookings.

On our call I noticed that half of the traffic was still going to the two weaker openers. Each variant already had hundreds of conversations behind it, so the result was solid. With one click I switched them off. Half of the traffic had been working at half the efficiency of the winner, so the whole funnel got about 25% more effective overnight.

So look at two columns: replies, to know whether the test is ready, and conversions from welcome to booked call, to pick the winner.

Keep the winner, test around it: 80/20

Once you have a winner, don’t keep sending half of your traffic to experiments. You built that opener over weeks. If something works, don’t break it.

Instead:

• 80% of traffic goes to the winner and small variations of it. Duplicate “you there?” and add a heart in front, or end it with three question marks. Take “can I ask you something?” and drop “can I”: “I have a question”. It looks like nothing, but small changes like these can be worth real money over a year. When one variation beats the original, make variations of the variation. The 80% can be split, for example 50, 20 and 10.
• 20% goes to completely new ideas. Think of it as rolling a die. Ten new ideas may lose, and the eleventh or the twentieth beats everything you had.

The data has one caution here: emoji usually lost in the study above. That doesn’t mean your heart variation will lose, but test it like anything else instead of assuming.

Then put a weekly slot in your calendar to review the tests: turn off variants with 200+ replies that clearly lose, and add one or two new ones. Stop after three weeks and you leave money on the table every month after.

Tell the AI what to do after “you there?”

A common worry: if the opener is “you there?”, won’t the AI setter get confused when someone replies “yes?”

It will usually handle it. But I recommend making the reply to the opener the first, scripted step of your sales process. In your AI appointment setter instructions, write a short critical rule for it. Words like “critical”, “always” and “never” make the AI follow a rule even when something else seems more natural.

Write the rule for openers “in the style of” your variants, not for an exact match, so you don’t have to update it every time you test a new variation. For example:

Critical rule: always use the profile analysis to open the conversation and build natural, casual small talk. Whenever someone replies to your openers in the style of “you there?”, “can I ask you something?” or similar hooks, give one casual, non-obvious compliment that doesn’t sound forced, and always start it with “by the way”. Then move to your first discovery question.

Add examples, because they work better than descriptions:

by the way, I see you’re into kitesurfing. I’m actually writing to ask whether…

by the way, do you play golf competitively or just for fun?

Setor reads the lead’s profile the way a person would: the bio, the posts and, on public profiles, the highlights, plus whether it’s a business or a private account and the tone they write in. That’s what lets the compliment be specific to the person. It’s one friendly, specific observation, then straight to the point. You can also add a condition: if it’s a business profile with 10,000+ followers that probably has an offer, ask how they get clients on Instagram; if it’s unclear, go more carefully.

FAQ

What is a good reply rate for an Instagram welcome message? Most of our clients start at 10 to 20%, and a few weeks of A/B tests usually take them to 30 to 40%. Across active Setor accounts, the median is 26.9% and the top tenth gets 44.1%. It varies by niche, so not breaking 30% doesn’t mean something is wrong.

How long should a DM opener be? Short. Openers of up to 8 words beat longer ones on 81% of the accounts that ran both. Aim for one line that looks typed by hand.

How many conversations do I need before picking a winner? About 200 replies per variant, counting started conversations, not sent messages. At a 40% reply rate that’s around 500 sends per variant.

Should the opener ask about their goals right away? No. The opener has one job: get a reaction. Questions, lead magnets and the offer belong in the next steps, which you script in the AI’s instructions.

Where the numbers come from

• Opener study: Setor database, read on November 2, 2026. 292 welcome versions with at least 300 sends each, 91 accounts, 723,857 sends. Each trait is compared only on accounts that ran versions with and without it. Reply rate without reminders. Measured mostly on Polish-speaking accounts.
• Reply rate ranges (26.9%, 34.5%, 44.1%): the same database, last 90 days, 50 accounts with at least 500 sends.
• Starting points of 10 to 20%, the 9 to 10% cases, “you there?” at 65% then 50%, the script drop from 15% to 7.5%, the fitness client’s 1.5 vs 3 bookings per 100 welcomes and the 200-reply rule: my own practice with Setor clients and my own profile, not a study.

Want every new follower to get an opener you can test? Start now and set up your first A/B test. Once more people reply and book, read how to reduce no-shows on sales calls.