Cold email sequences that convert Instagram-extracted B2B leads

A clean Instagram-extracted list with strong deliverability is necessary but not sufficient. Plenty of teams land a 6,000-row file with verified emails, warm the domains correctly, and still see reply rates under 1%. The list was not the problem. The sequence was — wrong opener, wrong personalisation, wrong cadence, and a CTA that asked for too much too soon. This post is the sequence-design companion to our deliverability playbook: deliverability gets the message into the inbox, sequence design gets the reply.

The argument here is specific to Instagram-sourced contacts, because they behave differently from a list scraped off a generic B2B directory. The recipient published an email on a profile they actively curate. They have a bio, a follower count, a posting style, and a reason they were on the seed account or hashtag you extracted from. All of that is signal, and a sequence that ignores it reads exactly like the 40 other cold emails in the inbox that day.

Why Instagram-sourced contacts need a different sequence

A contact pulled from a B2B data vendor is anonymous in the way that matters: you know the company and the job title, but you have no behavioural context. An Instagram-extracted contact carries three pieces of context that a directory record never does.

• A reason they were extracted. They followed a specific competitor, engaged with a specific niche post, or matched a specific bio keyword. That is a topic they care about right now, not a static firmographic attribute.
• A self-authored bio. The recipient chose how to describe themselves in roughly 150 characters. That line tells you what they want to be known for — and gives you a non-generic opener.
• A visible audience. Follower count, engagement style, and posting cadence indicate whether this is a solo founder, a small agency, or a brand with a content team. The pitch that lands differs across those.

The mistake is to flatten all of this back into “Hi {first_name}, I help companies like {company} do X.” That throws away the only advantage the source gave you. The whole point of an Instagram extraction over a bought list is that the source signal is the personalisation hook — use it.

Sequence architecture: four touches, not seven

The instinct on a fresh list is to maximise touches. Resist it. For Instagram-sourced B2B contacts, the working structure is four emails across 14–16 days. More than that and you are training the recipient — and the inbox provider — to treat you as noise.

Touch	Day	Job of this email
1	0	Earn the open and the reply with relevance — lead with the source signal
2	3	Add one concrete proof point; restate the ask more softly
3	7	Pattern-interrupt — short, different angle, easy yes
4	14	Explicit break-up; lowest-friction CTA

Four touches at this spacing typically pull 70–85% of the total replies a longer sequence would, at a fraction of the spam-complaint risk. The marginal fifth, sixth, and seventh touches mostly generate unsubscribes and complaints, both of which damage the sending reputation you spent two weeks warming. The cadence also respects the suppression discipline that matters if you are running the same list as a paid custom audience in parallel — fewer email touches means cleaner cross-channel frequency.

Touch 1: the opener that uses the source signal

The first email does one job: prove in the first line that this is not a blast. The lever is the extraction reason.

If the contact was extracted as a follower of a competitor, the opener references the category, not the competitor by name (naming the competitor reads as creepy and invites a defensive reply):

“You’re clearly deep in the {niche} world — most of the people I reach who are pay close attention to {category outcome}.”

If the contact came from a hashtag or location extraction, the opener references the topic or place — the same source signals we describe attaching at extraction time in the hashtag and location targeting post:

“Saw you’re active in the {city} {niche} scene — quick question about how you’re handling {specific problem}.”

If the contact matched a bio keyword, the opener reflects their own self-description back, lightly:

“Your bio says {paraphrased bio claim} — that’s exactly the kind of operator this is relevant for.”

Three structural rules for touch 1:

1. No pitch in the first two sentences. The first job is to earn the read, not to sell. The offer comes after the recipient has registered that you actually know who they are.
2. One question, not a calendar link. Asking for a 30-minute call in email one is the single most common reason Instagram-sourced sequences underperform. Ask a question that can be answered in one line.
3. Subject line under 5 words, lowercase, no brand name. “quick question about {topic}” outperforms “Scraphex — boost your leads 3x” on Instagram-sourced lists by a wide margin, because the audience is allergic to anything that smells like a SaaS funnel.

Expected behaviour: a well-built touch 1 on a properly filtered Spanish B2B list lands a 35–55% open rate and a 1.5–3% reply rate on its own. The rest of the sequence is there to roughly double the cumulative reply rate, not to do the heavy lifting alone.

Bio-based personalisation at scale without writing 6,000 emails

The objection to source-signal personalisation is always the same: it does not scale to a 6,000-row list. It does, if the deliverable is structured for it. The trick is to personalise the first line only, and to drive that first line from columns that already exist in a well-built extraction.

A Scraphex-style deliverable carries a source_signal column (which competitor / hashtag / keyword the row came from), a bio_snippet column, and a niche column. A merge template with three or four first-line variants keyed off the source_signal type produces a different, relevant opener for every row without a human writing each one.

• Build 3–4 opener templates, one per source-signal type (competitor-follower, hashtag, location, bio-keyword).
• Within each template, slot the niche and the bio paraphrase from the row’s columns.
• Keep the body — touches 1 through 4 after the first line — identical across the list. The personalisation budget belongs entirely in the first line, where it does the most work per character.

This is the difference between a list that is enriched for sequencing and a raw email dump. A raw dump forces you to choose between generic copy (low reply rate) and manual personalisation (does not scale). A structured deliverable lets you have relevance at scale, which is the only version of cold email that works on this audience in 2026.

Touches 2–4: proof, interrupt, break-up

Touch 2 (day 3) — one proof point. Reply in-thread to touch 1 so the original context is visible. Add exactly one concrete, specific proof point — a number, a named result pattern, a before/after. Restate the ask, softer than touch 1. Do not re-explain who you are; the recipient either remembers or will scroll up. Two short paragraphs, maximum.

Touch 3 (day 7) — pattern interrupt. Change shape entirely. If touches 1 and 2 were paragraphs, touch 3 is one line plus a question. The goal is to catch the recipient who skimmed past the first two because they looked like every other thread. “Still worth a conversation, or is {problem} not a priority this quarter?” gives an easy out, which paradoxically lifts replies — people answer “not a priority right now” and that is a real, usable signal.

Touch 4 (day 14) — the break-up. State plainly that this is the last email. Lower the CTA to its floor: not a call, not even a question, just “reply with a single word if you’d ever want me to circle back.” Break-up emails reliably produce 15–30% of a sequence’s total replies, because the explicit ending removes the obligation the recipient felt to do something effortful. Honour it: a contact who does not reply to touch 4 comes off the active sequence and goes to the suppression list, not back into another cadence next month.

Timing and volume that protect the domain

Sequence design and deliverability are not separable. A perfect sequence sent at the wrong volume burns the domain that makes the sequence possible.

• Ramp, do not blast. A freshly warmed domain sends 20–40 cold emails per day per mailbox, climbing 10–20% weekly. A 6,000-row list across, say, four mailboxes is a multi-week send, not a one-day event. Plan the calendar before importing the list.
• Send windows matter for this audience. Spanish B2B founders and agency owners read email Tuesday–Thursday, 8–10am and 4–6pm local. Avoid Monday morning (inbox triage) and Friday afternoon.
• Segment the send by quality tier. Send to the highest-confidence rows first (verified email + phone present + on-niche bio). Their reply and complaint behaviour tells you whether the copy is working before you commit the rest of the list. If touch 1 to the top 600 rows draws complaints, stop and fix the copy — do not push the remaining 5,400.

The full ramp-and-window discipline lives in the deliverability playbook; the point here is that the sequence calendar and the deliverability calendar are the same calendar.

Reply handling: the part most teams under-build

A sequence that generates replies and then fumbles them wastes the entire list. Replies from Instagram-sourced contacts cluster into four types, and each needs a pre-written, fast response.

• Interested (“tell me more”). Respond within working hours, not days. Move to the real ask — a call or a sample — now that interest is established. Speed here is the single biggest lever on conversion.
• Objection (“how did you get my email?”). This one is specific to extracted lists and must be answered honestly and calmly: the email was publicly published on their Instagram profile, you process it under legitimate interest, and they can opt out in one reply. The defensible version of this answer is exactly the posture we lay out in the GDPR and lawful-basis post. A confident, honest reply converts a meaningful share of these into conversations; a defensive or evasive one guarantees a complaint.
• Not now (“circle back in Q3”). Tag it, suppress from the current sequence, and actually circle back when they said. This is the highest-ROI bucket over a year and the most commonly dropped.
• Opt-out. Process immediately, across email and any paid audience the contact sits in. Confirm in one line. A clean, instant opt-out is both a legal requirement and a reputation protector.

What the data layer has to provide for any of this to work

Every technique above depends on the deliverable carrying more than an email column. The sequence is only as good as the columns it can merge from.

• Source-signal column — drives the first-line personalisation that makes touch 1 land.
• Bio snippet — the raw material for the self-description opener.
• Niche / category tag — lets you run different copy per segment instead of one generic body.
• Verification status and phone presence — defines the quality tiers you send in order.
• A schema that supports suppression — so opt-outs and replies leave the active sequence cleanly.

A list that is only an email column forces generic copy, which is why bought lists convert so poorly regardless of deliverability. A list extracted with the source signal preserved is a list you can actually sequence.

Test the sequence on a real sample first

The cheapest way to find out whether this approach fits your offer is to run it small before committing a full extraction.

Request a free sample and we will deliver a 50-row segment from a niche of your choice with the full column set — verified email, phone where public, source-signal attribution, bio snippet, and niche tag. That is enough to build your four-touch sequence, run it against a real segment of your target audience, and read the reply behaviour before scaling to a full 6,000-row campaign. The reply rate on a well-built sample sequence is usually within a couple of points of what the full list produces, which makes it a clean test of both the copy and the fit before you spend on volume.