Most Instagram DM templates fail for the same reason chain letters fail: they were written for someone else's customer. Paste forty words that worked for a supplement brand into a jewellery DM and it reads as exactly what it is — a script, arriving where a person should be.
So treat what follows as shapes rather than copy. Every one of these is the same three-part move — answer the actual question, be specific enough to be useful, leave a door open — applied to the eight situations that make up most of a DTC inbox. The words are illustrative; the structure is the part that transfers.
Each pattern below has the situation, the shape, why it works, and the way it goes wrong.
The rule underneath all eight
Before the list: every good DM reply does three things in this order, and the order is the whole trick.
Answer the question. Say something only your business could say. Ask one thing back.
Skip step one and you sound like a bot. Skip step two and you sound like any brand. Skip step three and the conversation ends, which mechanically matters — a customer's reply is what opens Meta's 24-hour messaging window, and that window is where selling actually happens.
1. The answer-first reply
Situation: “is this in stock?” — the most common message you will ever get.
The shape: “[Yes/no, plus the specific state]. [One detail that helps]. [One question].” → “It is — just restocked in the olive. What size are you usually?”
Why it works: the first line is the answer. No greeting, no thanks-for-reaching-out, no delay. The customer's question is resolved before they finish reading the first sentence, which is the single loudest human signal available to you.
How it fails: hedging. “Let me check on that for you!” is a reply with no information that costs the same as a real answer. If you have to check, say what you are checking and when you will be back.
2. The one-question qualifier
Situation: a vague opener — “how much?”, “info?” — where answering directly would be a guess.
The shape: “[Partial answer]. [One question that makes the next answer exact].” → “They start at $98 — are you after the wide-leg or the straight?”
Why it works: it gives before it asks. A bare “which product?” makes the customer do the work; leading with the price range earns the right to ask. One question, never two — a two-question reply is an interview, and people leave interviews.
How it fails: asking for the sale before knowing what they want. Nobody buys from a question mark.
3. The specific reassurance
Situation: the fit or quality doubt. “Do these run small?” This is the last thing standing between a browser and a buyer.
The shape: “[The real fit note]. [The de-risk]. [The ask].” → “They run true to size, so your usual M is right — the fit's a relaxed leg, not slim. Returns are free for 30 days if it's not perfect. Want the link?”
Why it works: it answers with a fact rather than a deflection, removes the risk, and asks. “Check our size guide” does none of the three — it hands the customer homework at the exact moment they were ready to buy, and homework at the point of impulse means the purchase dies.
How it fails: when the fit note is invented. A confident wrong answer here costs a return, a refund and a review. This pattern only works if the specific is true.
4. The comment-to-DM opener
Situation: they commented your campaign word and you get exactly one private reply. Meta's private reply rules allow one message per comment, within seven days of it being posted — that is the whole budget.
The shape: “[What they asked for, delivered]. [The fact that helps]. [The door].” → “Got you — that's the olive wide-leg, $98 and in stock in your size. Here's the link, or want me to check the fit first?”
Why it works: it spends the single message on an answer and a door rather than a bare URL. The reply it earns is what opens the 24-hour window; a link dump ends the exchange while you still had a day of usefulness available.
How it fails: the vending-machine version — comment in, link out, silence. It converts a fraction of what it could, and the send count looks identical either way. The comment-to-DM playbook covers the campaign strategy around this one.
5. The stall nudge
Situation: “ok cool thanks” — and then nothing. Not a no. A parked purchase, most of the way there.
The shape: “[Low-pressure offer of help]. [A gentle reason to act].” → “Anytime! Want me to hold one in your size for a bit? They've been moving fast this week.”
Why it works: it reads as service, not pressure, and it gives them permission to say yes to something small. Almost no brand sends it, because no human is watching the thread twenty minutes later — which is the entire reason it is worth automating.
How it fails: repetition. One nudge is a helpful shop assistant. Three is a stalker, and the second one is where the goodwill goes.
6. The out-of-stock save
Situation: they want the thing and you do not have it. Most brands treat this as the end of a conversation. It is the middle of one.
The shape: “[Honest no]. [The nearest real alternative or a date]. [The ask].” → “Not in the M right now — the L runs close to a relaxed M if you like it roomy, or I can message you the second it's back. Which would you rather?”
Why it works: it converts a dead end into two live options, and the “tell me when it's back” branch captures intent you would otherwise lose entirely.
How it fails: pretending. Substituting a product that is not actually similar to make a sale today costs you the customer tomorrow.
7. The honest handoff
Situation: a question your automation genuinely cannot answer — a custom order, a damaged parcel, anything with feelings attached.
The shape: “[Admit the gap plainly]. [Name what happens next]. [Give a time].” → “Good question — not something I want to guess at. Passing this to Sam now, you'll hear back this morning.”
Why it works: admitting a gap is the most human sentence in customer service and the one line a keyword bot can never say. It also beats the alternative, which is a fallback that pretends the question was handled.
How it fails: when nobody is actually tapped in. A handoff that goes nowhere is worse than no handoff, because you promised.
8. The price answer that does not flinch
Situation: “how much?” on something expensive.
The shape: “[The number, first]. [What justifies it, concretely]. [The ask].” → “$240. It's the heavyweight wool — the same jacket most places do in a blend for $180. Want the fit details?”
Why it works: leading with the number signals confidence; burying it signals apology. The justification has to be a fact, not an adjective — “premium quality” is what everyone says, “heavyweight wool, not a blend” is what only you can say.
How it fails: leading with a discount. Answering a price question with a coupon teaches your customers to always ask twice, and trains the audience you least want.
The eight at a glance
| Pattern | Fires when | The mistake it replaces |
|---|---|---|
| Answer-first reply | Any direct question | “Thanks for reaching out!” |
| One-question qualifier | A vague opener | Answering the wrong product |
| Specific reassurance | The fit / quality doubt | “Check our size guide” |
| Comment-to-DM opener | Your campaign word, one message only | A bare link |
| Stall nudge | “ok cool thanks” + silence | Nothing at all |
| Out-of-stock save | You do not have it | “Sorry, sold out!” |
| Honest handoff | You genuinely do not know | A confident guess |
| Unflinching price | “how much?” on a big number | Leading with a discount |
Build this one first
If you only ever set up one: the stall nudge. Every other pattern on this list improves a conversation you were already having. The nudge recovers conversations that are already over — threads where the customer got their answer, said thanks, and drifted. Those are the closest to money and the least contested, because nobody sends it.
Making Instagram DM templates sound like you
Good Instagram DM templates are not words you paste — they are a structure you internalise: answer, be specific, leave a door. The words should change every time, because the customer does. Anything you send forty times identically will be screenshotted on the forty-first.
Take the eight shapes, rewrite them in your own voice against your last twenty real DMs, and notice which ones you cannot fill in because you do not know the answer yourself — that gap is the actual work. If you would rather see the shapes running against a live catalog than maintain them by hand, that is what HyperDM does across use cases; the first 50 conversations a month are free on pricing. And for the bad versions of all eight, the DM teardown is the companion piece.
Run the automations. Let the AI close.
Comment-to-DM, keyword replies, follow gates — set up in minutes, then the AI answers from your real catalog. Free on one channel, 50 conversations a month, no card.