Reply.io: Add Inbox Placement Check to Your Sequences

Reply.io is a multi-channel outbound platform. It runs cold email, LinkedIn touches, calls, and SMS in a single cadence called a sequence. For cold email specifically, that means every recipient in a sequence receives four, six, or eight scheduled emails over several weeks. Each of those emails has its own subject line, its own body, and its own risk of landing in spam. If stage one hits the inbox and stage three hits Promotions, the whole sequence quietly dies and you only notice in the reply rate weeks later.

Seed testing fixes the blind spot. You add a small number of dedicated seed mailboxes to the sequence contact list the same way you would add a prospect. They receive every stage on the same schedule as real recipients. You check where each stage lands and you have a per-stage placement map for the full cadence. This article walks through how to set that up inside Reply.io, what to watch for with multi-channel steps, and how to interpret the results.

Why per-stage placement matters in Reply.io

Sequence tools behave differently from single-send ESPs. A broadcast email either hits or it doesn't. A sequence has a compounding risk: stage one may pass, but stage two has a different subject line, a different first line, and arrives at a different time of day. Spam filters also score follow-ups more harshly when the previous message was not opened. If a recipient never engages with stage one, stage two is more likely to go to Promotions or spam at Gmail. Your reply rate collapses and you blame copy when the real cause is placement drift across stages.

Reply.io surfaces open and reply metrics per step, but opens are unreliable in 2026 because Apple Mail Privacy Protection and Gmail image proxies inflate them. Placement data cannot be faked that way. Seeds tell you exactly which folder the stage arrived in, which is the number that actually predicts deliverability.

Reply.io sequence structure, quickly

A Reply.io sequence is a list of steps. A step is one of: email, LinkedIn action, call task, or SMS. Email steps have a delay (for example "Wait 3 days after previous step"), a subject template, a body template with Liquid-style variables, and sending windows. The platform rotates sending across your connected mailboxes if you have more than one and throttles per-mailbox per-day.

The typical cold-email shape looks like this:

• Step 1 — Email, day 0, unique subject
• Step 2 — Email, day 3, reply-in-thread
• Step 3 — LinkedIn connect, day 5
• Step 4 — Email, day 7, new subject
• Step 5 — Email, day 12, short bump
• Step 6 — Email, day 18, breakup

Four of those six steps are email. You want placement data for all four, measured from the same sequence run with the same mailbox, the same warm-up state, and the same headers.

How to add seeds to a Reply.io sequence

1. Generate a fresh seed list. Open the Inbox Check tool on the homepage, copy the 20+ seed addresses it produces for this test. Each run gives you fresh mailboxes across Gmail, Outlook, Yahoo, Mail.ru, Yandex, ProtonMail and more. Don't reuse a seed list across weeks — engagement signals decay and filters learn the pattern.
2. Create a test contact list. In Reply.io go to People → Create list, name it seed-test-2026-10-28. Import the seed addresses as CSV. Use plausible first/last name fields so your personalization variables don't render as "Hi ," with a missing name — that alone can tank placement.
3. Attach the list to your live sequence. Open the sequence and click Contacts → Add → pick your seed list. You can do this alongside real prospects or in a mirror sequence. A mirror sequence is cleaner: identical steps, identical mailboxes, identical schedule, only the contact list differs.
4. Let all steps fire. Don't pause after step one. The whole point is per-stage placement, which you only see if every step reaches the seeds.
5. Check each seed mailbox per step. Record Inbox / Promotions / Spam / missing for every stage per provider.

Mirror sequence vs. shared list

A shared list (seeds added to the live sequence) is faster but adds noise to your campaign stats. A mirror sequence isolates the test. For anything above ~500 live prospects, mirror it. For a small founder-led test under 100 prospects, a shared list is fine.

Multi-channel quirks that affect placement

Reply.io lets you mix email with LinkedIn and call tasks. The LinkedIn and call steps don't send email, but they affect sequence pacing and the day-gap between email steps. That pacing matters for placement in two ways. First, Gmail weights recency of previous engagement. If step 2 comes three days after step 1, deliverability is usually fine. If there's a 10-day gap because of a LinkedIn + call break in the middle, the follow-up may be treated more like a cold initial message and filtered harder.

Second, Reply.io reply-in-thread behavior keeps the same Message-ID chain. Threaded follow-ups benefit from the first message's engagement signals if the recipient opened or replied. If nobody engaged, threading does not help and can actually hurt — you're stacking on a cold thread. Seeds don't open or reply by default, so your seed data reflects the worst-case "no engagement" path, which is also the most common real-world path. That's a feature, not a bug.

Reading the results

Build a small table keyed on step × provider. For each cell, log the landing folder:

Step   Gmail   Outlook   Yahoo    Mail.ru   ProtonMail
1      Inbox   Inbox     Inbox    Inbox     Inbox
2      Inbox   Inbox     Promo    Inbox     Inbox
3      Promo   Inbox     Spam     Inbox     Inbox
4      Spam    Junk      Spam     Spam      Inbox

A clean sequence looks like the first row for every step. Degradation down the table is normal — that's fatigue. Fast degradation (step 3 already in spam at two providers) means one of three things: the subject line of step 3 is the problem, the gap before step 3 is too large or too small, or the body has a hard pattern (lots of links, spammy phrase, attachment).

What to fix first

• If Gmail specifically drops but Outlook holds: rework subject line and first-line personalization for step 3.
• If Outlook drops but Gmail holds: check SPF alignment, the URL reputation of any links (especially tracking redirects), and List-Unsubscribe presence.
• If Yahoo and Mail.ru drop earlier than Gmail/Outlook: likely DKIM/DMARC alignment issue, or your sending domain is too young.
• If ProtonMail is the only clean provider across every step: that's actually a warning. ProtonMail is permissive and won't catch problems. Don't use it as a positive signal.

Cadence hygiene for Reply.io sequences

Run a seed test on every new sequence before adding real prospects. Run a seed test on every copy change — a subject line rewrite can flip placement overnight. Run a seed test when you add a new sending mailbox to the rotation. And re-run a seed test monthly on live sequences to catch reputation drift on the sending domain.

Keep a spreadsheet of seed results per sequence per date. Over time you will see which sequences degrade fastest and which sending mailboxes hold placement longest. That's the data that tells you when to retire a sequence, swap a mailbox, or warm a new domain.

FAQ