Aller au contenu
NouveauDécouvrir comment

Passer à l'échelle votre activité de cold email

Développez l'e-mail à froid sans perdre en qualité. Faites grandir ensemble boîtes, données, messages et gestion des réponses, et sachez quand ralentir.

A campaign that works for fifty contacts a week can fall apart at five hundred. Volume exposes every weak spot: thin research, tired mailboxes, slow replies, messy lists. Scaling well is not about sending more email. It is about growing each part of the system in step, so quality and reputation hold while the numbers rise. This chapter shows what to grow, in what order, and the warning signs that tell you to slow down.

Prove it before you grow it

Scaling something that does not work just produces more of the thing that does not work, faster, and burns through your list and your sender reputation on the way. Before you add volume, make sure the small version is healthy. That means a segment that replies, a trigger that holds up, a message people answer about the substance, and a team that answers interested replies the same day.

Ready to scale?If any of these is still shaky, fix it at the current size first. Problems only get harder to see at volume.
Ready to scale?: Steady interested replies, Low bounces and complaints, A repeatable research process, Fast answers, A written playbook0/5done

0 of 5 checked. Steady interested replies: Week after week from the same segment, not one lucky batch.

The four things that must grow together

Every cold email program rests on four parts. Grow one without the others and something breaks. More contacts without more mailboxes means each mailbox sends too much. More mailboxes without better research means more generic email. More email without more reply capacity means warm leads wait for days.

The layers of a program that scalesEach layer depends on the one below it. Pick a layer to see what it needs as volume grows.
The layers of a program that scales: Infrastructure, Data, Messaging, Reply handling01Infrastructure02Data03Messaging04Reply handling

Infrastructure. Domains, mailboxes, authentication and warm-up. Sets how much you can send safely.

Infrastructure: more mailboxes, not louder ones

Each mailbox can only send so much before providers start to treat it with suspicion. The safe way to send more is to spread the volume across more mailboxes, each sending a modest, steady amount, rather than pushing one mailbox harder. Chapter 4 covers daily limits and warm-up in detail. At scale, the same rules apply, multiplied.

  1. Use separate sending domains

    Send cold email from domains that are clearly yours but separate from your main company domain. If one has trouble, your everyday email and your main domain are protected.

  2. Set up every domain properly

    SPF, DKIM and DMARC on each one, as Chapter 3 describes. Every new domain needs the same care as the first.

  3. Warm up each new mailbox

    Start low and raise volume slowly over weeks. A new mailbox sending at full volume on day one looks like a spammer.

  4. Keep a modest daily limit per mailbox

    Grow by adding mailboxes, not by raising each one's limit. Steady and boring is what inbox providers trust.

  5. Add capacity ahead of need

    Warm-up takes weeks, so start new mailboxes before you need them, not on the day the next campaign launches.

More domains is not a way around the rules

Spreading volume across mailboxes is about keeping each one at a natural pace. It does not make bad email acceptable. If people mark your messages as spam, more domains just spread the damage further. Fix the message and the list first.

Data: keep the supply clean

At small scale you can hand-check every contact. At larger scale you cannot, so the process has to do the checking for you. Write down exactly what makes a contact fit: industry, size, role, location and the trigger. Build lists from the same kinds of public, relevant sources each time, and record the source for every contact. Verify addresses before sending, and remove anyone already on your suppression list before they reach a sequence.

  • Watch for segment exhaustion. Every niche has a limit. When the best-fit contacts run out, quality drops even if the numbers look fine. Find the next segment before that happens.
  • Refresh triggers often. A trigger that was true last month may not be true now. Job ads close, posts age, news goes stale.
  • Do not lower the bar to fill the quota. A smaller list of good fits beats a larger list of maybes. Weak contacts bring weak replies and more complaints.
  • Keep one source of truth. One record per contact and per company, so two people never write to the same buyer with different messages.

Messaging: specific at any volume

The hardest thing to keep as you grow is the feeling that the email was written for one reader. The answer is structure. Keep the parts of the email that should stay the same, the bridge, the offer and the ask, fixed in a template you have tested. Let research fill only the parts that should change, the opening line and the trigger. That way each email stays specific, and you can still review drafts quickly.

Part of the emailAt small scaleAt larger scale
Opening lineWritten by hand per contactDrafted from a research note, reviewed by a person
TriggerFound by handFound by a defined research step, contacts without one dropped
Bridge and offerRewritten oftenFixed per segment in a tested template
AskVariesOne small, tested ask per campaign
Follow-upsWritten as you goPlanned sequence, each step adding something new

AI can help a lot here, as Chapter 11 explains, but only with a person reviewing every draft. At volume, review is not a bottleneck to remove. It is the step that keeps hundreds of emails from carrying the same mistake.

Reply handling: the part people forget

If a campaign works, more volume means more replies. That is the point, but it is also where many programs quietly fail. Interested replies wait for days. Referrals get lost. Opt-outs are missed in a busy inbox. Plan reply capacity before you raise volume, not after.

  1. Bring replies into one place

    Across every mailbox, so nobody has to check ten inboxes.

  2. Sort them every day

    Interested, not now, referral, not interested, opt-out, automatic. The labels from Chapter 12.

  3. Give every reply an owner

    A named person answers each interested reply the same working day.

  4. Act on opt-outs at once

    Remove the person from every campaign and add them to the suppression list, as Chapter 13 describes.

  5. Schedule the "not now" replies

    Note the date they gave and follow up then, with a reason.

Grow in steps, not jumps

Raise volume in steady steps, and hold each step long enough to see what it does to your numbers. A sudden jump makes providers suspicious and hides cause and effect. If something goes wrong after you doubled everything at once, you will not know which part broke.

A calm way to growAn example rhythm. The exact pace depends on your mailboxes and your market. Pick a stage to see what to check.
A calm way to grow: Weeks 1 to 4, Weeks 5 to 8, Weeks 9 to 12, After thatWeeks 1 to 4Weeks 5 to 8Weeks 9 to 12After that

Weeks 1 to 4. Run the proven campaign at its current size. Write the playbook. Start warming up new mailboxes.

Warning signs to slow down

SignWhat it usually meansWhat to do
Bounces risingThe data source is weakening, or lists are going stalePause the affected list, verify again, find better sources
Spam complaints risingMessages are reaching people who did not expect themTighten the segment and trigger before sending more
Replies falling, opens unclearPossibly a delivery problemCheck placement with seed mailboxes and Postmaster Tools before rewriting
More "who are you?" repliesResearch or relevance is slippingReview a sample of recent emails and their notes
Interested replies waitingReply capacity has fallen behindSlow sending until answers are back to same day
Unsubscribes jump at one stepThat step irritates peopleRewrite or remove the step

Building a team around it

Past a certain size, one person cannot research, write, review, send and answer everything. Splitting the work helps, as long as everyone follows the same playbook. Common roles are a researcher who builds lists and notes, a writer or reviewer who owns templates and approves drafts, and a closer who handles interested replies and meetings. In small teams, one person may hold two of these. What matters is that each part has a clear owner.

  • Agree who owns each account before a campaign starts, so two people never write to the same buyer.
  • Share one suppression list across the whole team and every tool.
  • Review together every week. The fifteen-minute check from Chapter 12, with everyone in the room.
  • Write things down. Every test result, every lesson, every template change. A playbook is what lets new teammates start well.

Scaling across markets

A new country is a new segment, not just a longer list. Language, tone, business customs and the law all change. Treat each new market as its own small test. Write in the reader's language if your team can answer in it. Check the rules for that country before you send, as Chapter 13 explains. Time sends to the reader's working hours, not yours. And keep separate results per market, so a strong home market does not hide a weak new one.

What to automate, and what to keep human

Scaling tempts you to automate everything. Some steps should be automated, because people do them badly when they are repetitive. Others should stay with a person, because they are where trust and judgment live. Drawing the line clearly keeps you fast without making you careless.

AutomateKeep human
Sending at a steady pace within each mailbox's limitDeciding who to write to and why
Stopping a sequence when someone repliesApproving every AI-drafted email before it goes out
Applying opt-outs across every campaignAnswering interested replies
Sorting replies into labelsReading a sample of replies each week for patterns
Checking new lists against the suppression listChoosing the next segment and the next test
Checking DNS and authentication on new mailboxesLinkedIn steps and phone calls, done at a human pace

A useful test: if a mistake in this step would reach a reader as something rude, false or unwanted, keep a person in the loop. If a mistake would only slow you down, automate it and check it now and then.

Doing this in FirstSales

FirstSales lets you connect several mailboxes, sets a daily limit per mailbox, and checks DNS when you connect each one. Campaigns can use Manual, Template or AI drafts with research notes and a review queue, so emails stay specific as volume grows. Sequences stop when someone replies, opt-outs apply across every campaign, and replies are sorted so interested ones reach you first. The Copilot helps set up campaigns in plain language. You can explore a multi-mailbox setup with sample data in the demo.

Key takeaways

  • Prove a campaign at small scale before you add volume.
  • Grow infrastructure, data, messaging and reply handling together.
  • Send more by adding warmed-up mailboxes, not by pushing each one harder.
  • Keep emails specific with fixed templates plus reviewed research lines.
  • Raise volume in steps, and slow down at the first warning sign.
  • Give every part of the work a clear owner and one shared playbook.

Réponses rapides sur la masterclass

Combien d'emails à froid envoyer par jour ?

Commencez par 5-10 par boîte et par jour pour les nouveaux comptes. Après 2-3 semaines de warm-up, montez à 35-50 par boîte. Si vous devez envoyer 500+ emails par jour, utilisez 10-15 boîtes chauffées et répartissez les envois sur l'ensemble. Ne poussez jamais une seule boîte au-delà de 50 emails à froid par jour.

Quel est un bon taux de réponse en emailing à froid ?

Il n'existe pas de chiffre unique et honnête. Le taux de réponse dépend de votre marché, de votre liste, de votre offre et de la façon de compter les réponses, donc les chiffres d'autres campagnes se transposent rarement. Comparez chaque campagne à vos propres résultats passés pour le même type de public, et jugez sur les réponses intéressées plutôt que sur le total. Le chapitre 12 montre comment.

Quelle longueur pour un email à froid ?

Gardez vos emails à froid entre 50-125 mots. Certaines études suggèrent jusqu'à 150 mots pour certaines audiences B2B. Les emails plus courts sont toujours plus performants. S'il vous faut plus de 125 mots pour faire passer votre idée, votre message manque de précision.

HTML ou texte brut pour les emails à froid ?

Texte brut. Toujours. Les emails HTML avec logos, images et mises en page formatées déclenchent les filtres anti-spam et ressemblent à des emails marketing. Un email en texte brut, d'une personne à une autre, voilà à quoi devrait ressembler un email à froid. La seule exception est votre signature, qui peut inclure une mise en forme basique.

Combien de relances faut-il envoyer ?

En général quelques-unes, espacées de plusieurs jours, chacune apportant quelque chose de nouveau : un autre angle, une ressource utile ou une question simple. Arrêtez dès qu'une personne répond, et terminez par une clôture courte et polie. Le chapitre 8 traite du rythme et du contenu de chaque étape.

Masterclass cold email

Une prospection qui continue quand vous vous déconnectez.Leads, emails, relances et réponses, gérés.

Connectez des boîtes illimitées et laissez les agents IA trouver les leads, écrire chaque email et relancer. Ou envoyez vos propres modèles en mode manuel. À partir de $29/mois.

$1 pour 3 jours · Annulez quand vous voulez · Sans engagement

Installer avec l'IA

Votre agent de code installe la CLI et les skills FirstSales pour vous.

Prompt pour Claude Code

I'm setting up FirstSales (https://firstsales.io), an AI outbound sales platform: campaigns, contacts, a reply inbox, deliverability and lead signals. Help me get my workspace sending real, good emails end to end, from Claude Code.

Read these first so you know how FirstSales works:
- Developer docs and API: https://developer.firstsales.io
- CLI: https://www.npmjs.com/package/@firstsales.io/cli
- Skills for coding agents: https://github.com/firstsalesio/firstsales-skills (read README.md and GUIDE.md)

Setup
1. Install the CLI: npm install -g @firstsales.io/cli
2. Install the FirstSales skills:
   claude plugin marketplace add firstsalesio/firstsales-skills
   claude plugin install firstsales@firstsales
3. Ask me if I already have a FirstSales account. If not, send me to https://app.firstsales.io to sign up.
4. Ask me to create an API key under Settings → API → API Keys, with only the scopes we need.
   Never ask me to paste the key into this chat. Tell me to run this in my own terminal, so the key stays hidden:
   read -rs FS_KEY && firstsales auth login --api-key "$FS_KEY"; unset FS_KEY
5. Run `firstsales whoami` and tell me which workspace and organisation we are connected to.

Then ask me about my business: what I sell, who I sell to (roles, company size, countries), my best customers, and what I use today for outbound (tools, lists, mailboxes). Walk me through:
1. Teaching FirstSales my business: website, offer, proof points and case studies in the knowledge base, so the AI writes like someone who knows the product.
2. Connecting my sending mailboxes and checking my sending domain: SPF, DKIM, DMARC and warm-up. If anything is unhealthy, tell me the exact DNS record to add.
3. Bringing in contacts: import my CSV or CRM export into a clean list, remove duplicates and bad emails, and block domains I must never email (customers, competitors, my own).
4. Building my first campaign: the right list, senders, steps and timing. Write the sequence with me, show me the drafts, and do not launch until I say yes.
5. Setting up Signals if it fits: find people engaging with LinkedIn posts about topics my buyers care about, and turn them into leads.
6. Inviting my team and giving each person the right access.

Once campaigns run, show me how to work the reply inbox from here: which replies need an answer, draft responses in my voice, and move good conversations into deals. If I use Cal.com, connect it so follow-ups can offer my booking link (not the first email).

Throughout, as you learn about my business, proactively suggest and help create:
- Reply templates for the questions and objections my prospects are likely to raise, based on my product, site and codebase if you can see them.
- Campaign ideas and segments worth testing, with the reason for each.

Safety rules:
- Show every change as a dry run (--dry-run) first and wait for my yes. Never send emails, launch campaigns or delete anything without my yes.
- Tell me before anything that spends credits, and how many.

Be opinionated. If the docs give a sensible default, just use it. Don't make me choose between five options.