Outbound vs inbound sales — when to use which (and how to run both)

Outbound and inbound are not fast and slow versions of the same thing. Outbound is a metered system: you set a dial, and volume comes out in proportion to mailboxes, hours and headcount. Inbound is a measured system: demand arrives at a rate you can influence over months but cannot set this week. Almost every practical difference — the legal basis, the failure mode, the forecast you can honestly make, the person who has to move fast — follows from that one structural fact rather than from anything about content or cold email.

The one difference everything else follows from

Who initiated the contact. That is the whole distinction, and it is worth being pedantic about because so much downstream behaviour hangs off it.

When you initiate, you are the party who has to justify the contact. You carry the legal obligations, the deliverability risk, the pacing constraint and the burden of relevance. When they initiate, all four move: the legal picture relaxes, the deliverability risk mostly disappears because the recipient expects to hear from you, and the burden shifts from relevance to speed.

The words "inbound" and "outbound" describe channels only by accident. A referral you asked for is outbound. A reply to a newsletter someone subscribed to is inbound even though it is an email you sent. Sorting your motions by who moved first is more useful than sorting them by medium, and it is the sort that predicts what will go wrong.

Outbound is a metered system; inbound is a measured one

Here is the practical consequence. Outbound throughput is a product of things you own.

• Number of connected mailboxes, and the per-mailbox daily cap.
• Hours of rep time available for calls and LinkedIn work.
• Size of the addressable list that survives your ideal customer profile.

Multiply, subtract the compliance losses, and you have a forecast that is arithmetic rather than hope. If the number is too small, you can name which factor to change.

Inbound throughput is a product of things you do not own: search demand for your category, how many other people are answering the same question, and how a search engine or an answer engine ranks the results this month. You can invest in each of those, and the investment often pays. What you cannot do is decide on Monday that next week's inbound will be forty per cent higher.

That asymmetry is why the two motions belong to different planning horizons. Outbound is a quarter-level lever. Inbound is a year-level asset. Treating either one as the other is the most common planning mistake in a small go-to-market team.

What the law changes the moment you initiate

This section is jurisdiction-specific and worth reading slowly, because the rules for a cold message and a requested one genuinely differ.

In the United States, the Federal Trade Commission's CAN-SPAM Act compliance guide states that the law "covers all commercial messages", defined as any email whose primary purpose is "the commercial advertisement or promotion of a commercial product or service", and — the sentence that surprises people — "the law makes no exception for business-to-business email". CAN-SPAM does not require prior consent. It requires accurate headers, a non-deceptive subject line, identification as an advertisement, a valid physical postal address, a working opt-out, and that you honour that opt-out within 10 business days while keeping the mechanism able to process requests for at least 30 days after the message went out. The FTC's guide states each separate non-compliant email can carry penalties of up to $53,088.

In the United Kingdom the shape is different. The Information Commissioner's Office's business-to-business marketing guidance says corporate subscribers are bodies "with separate legal status (eg companies, limited liability partnerships, Scottish partnerships, and some government bodies)", and that "the main difference is that the rule on marketing by electronic mail (eg email or text message) doesn't apply to corporate subscribers". The trap is in the next sentence: "sole traders and other types of partnerships are classed as 'individual subscribers' and PECR treats them the same as individuals". A Scottish partnership is a corporate subscriber; an English one is not. And UK GDPR still applies regardless — the ICO says you "must tell them about this and have a lawful basis under the UK GDPR for the processing", and that business contacts "can object to your direct marketing". That guidance carries its own caveat as of this writing: the ICO notes it is under review because of changes made by the Data (Use and Access) Act.

Inbound sidesteps most of this. Someone who filled in your form gave you their details for the purpose you stated on the form. The obligations do not vanish — the privacy notice and the objection right survive — but the consent question is answered before you send anything.

The arithmetic of outbound capacity, worked

Numbers below are an illustrative worked example on Autocloz's shipped defaults, not a measured result from any customer.

Autocloz starts a newly connected mailbox at 40 sends a day, 10 an hour, with a 90-second minimum gap between sends. Ten mailboxes therefore give a hard ceiling of 400 messages a day. Follow-ups eat into that ceiling, because a step three going out today competes with a step one for the same capacity, so a steady-state sequence with four email touches spends roughly four sends per lead over its life.

Suppose you want 25 first touches a day, sustainably, on those ten mailboxes. Four touches per lead means about 100 sends a day at steady state, comfortably inside 400. Take a 2% positive-reply rate on a well-built list — plausible for cold B2B email, and not a promise — and 25 first touches a day is roughly 0.5 positive replies a day, or about 11 a month. If one in three positive replies becomes a meeting and one in four meetings becomes an opportunity, that is a shade under one opportunity a month per 25 daily first touches.

Now the sentence that makes this useful: if you need four opportunities a month, you need roughly 100 first touches a day, roughly 400 sends a day at steady state, and therefore roughly 40 mailboxes at the default cap — or fewer mailboxes with raised caps, which costs deliverability headroom instead of money. That is a real constraint you can plan against, and it is the sort of arithmetic the outbound stack cost calculator is for.

The complaint budget is the second ceiling and usually the binding one. Google asks senders of more than 5,000 messages a day to personal Gmail accounts to keep the spam rate reported in Postmaster Tools below 0.30%, and recommends staying below 0.10%. That is a per-message budget, and every additional touch spends some of it. How many follow-ups to send works that trade-off through properly.

Autocloz's free plan covers 5 users and 10 mailboxes with warmup and SPF, DKIM and DMARC monitoring included, which is enough to run the 25-a-day shape above — start free and see what your own list does before scaling the mailbox count.

Why the same arithmetic does not work for inbound

You can write the identical equation for inbound. Sessions times conversion rate times qualification rate. The equation is fine. The first term is the problem: you cannot set it, and its variance is much wider than an outbound send count's.

So inbound gets planned the other way round — you fix the conversion machinery and let the traffic term float. Concretely that means investing in the things that hold their value at any traffic level: a page that answers the question a buyer actually typed, a form that asks for the minimum, a routing rule that assigns an owner instantly, and a first response measured in minutes.

There is one more asymmetry worth naming. Outbound failure is loud — bounces, complaints, a sequence that stalls with a reason attached. Inbound failure is silent. A form that broke last Tuesday produces exactly the same absence of leads as a quiet week, and nothing in your CRM distinguishes the two. Anyone running inbound should have a synthetic submission test on a schedule for precisely this reason.

The failure mode each motion has, and how you diagnose it

They fail differently, and mixing up the diagnostics wastes weeks.

Outbound fails by over-contacting. The symptom is a rising complaint rate, falling delivery, or a channel account getting restricted. The diagnosis is per-mailbox and per-domain: look at sends per mailbox per day, the spam rate in Google Postmaster Tools, and the bounce classification split. Reducing bounce rate covers the list-hygiene half of that.

Outbound also fails quietly by stalling. A campaign that is running but sending nothing usually has a gate holding it — outside the sending window, no active mailbox, an hourly throttle. Autocloz stamps a last_skip_reason on every deferred enrolment for exactly this, so a parked step reads "Waiting: outside sending window" instead of a mute "in progress".

Inbound fails by under-responding. The symptom is a healthy form-fill count and a poor meeting count. The diagnosis is time-to-first-touch, measured from the submission timestamp rather than from when a rep opened the record, and segmented by hour of day so you can see the overnight gap.

Inbound also fails by mis-qualifying. A demand-generation page that attracts students, competitors and job seekers produces volume that looks like success and converts like noise. The fix is a qualifying field on the form, not a stricter rep.

Running both without letting one poison the other

Three rules, and they are all about shared state.

One suppression list, workspace-wide. In Autocloz, suppressions are keyed on the workspace and the address, with a scope of either email for one address or domain for a whole domain. Every email path consults the same helper — the campaign worker, campaign enrolment, the shared inbox's compose and reply, a test send and a manual lead send — so an unsubscribe from a newsletter also stops an outbound sequence to the same person. The non-email channels run off a separate do-not-contact list keyed on phone number or handle, so if you want a single exit across all five channels you have to add the person to both. If your two motions keep separate opt-out lists entirely, you will eventually email someone who opted out, and that is the single most expensive mistake in this whole area.

Stamp provenance at creation, never infer it later. Every lead row should carry where it came from, written when the row is created. Inferring provenance afterwards from which campaign touched a lead first is wrong the moment an inbound lead gets enrolled in an outbound sequence, which happens constantly.

Qualify per provenance. A demo request and a scraped contact should not clear the same bar to become an opportunity. Applying one bar to both either lets weak outbound leads through or holds strong inbound ones back. The distinction between a marketing-qualified and a sales-qualified lead exists precisely to make that explicit, and the difference between an MQL and an SQL is worth settling before you write the rule.

Practically, the surfaces that catch inbound need to be as owned as the outbound campaign is. Autocloz's booking pages and calendar sync are free with no per-seat fee, and its contact forms deliver a submission into a chosen Leads List so the enquiry lands somewhere a rule can act on it. That is the join point: inbound arrives in a list, and a list is something outbound machinery already understands.

What changes between one seller and ten

At one seller the two motions share a brain, so nothing needs coordinating. The whole design problem is time allocation, and the honest answer is to block outbound work rather than fitting it around inbound, because inbound interrupts and outbound does not.

At ten, three things break at once.

Ownership becomes ambiguous. An inbound enquiry from an account someone is already sequencing needs a rule, decided in advance, about who keeps it. The usual rule — the existing owner keeps it, and the enquiry is logged on their timeline — is right more often than the alternative, because the buyer has already had a conversation with that rep.

Attribution becomes contested. Two motions touched the same account, and both teams will claim it. Settle this by writing the rule down before anyone has an incentive, and prefer a rule that is boring and checkable over one that is fair and arguable.

Visibility becomes a permission question. Ten people sharing one lead database need row-level scope, not just module access. Autocloz resolves four scopes per module and action — off, own, team and all — with team meaning rows owned by anyone on the holder's team, and owner and admin roles resolving to all. That is what team collaboration with roles and scopes is for, and it is worth configuring before the tenth person joins rather than after.

If you are weighing whether an all-in-one workspace or separate best-of-breed tools handles this better, the all-in-one versus best-of-breed decision lays out the trade-off, and the Autocloz and HubSpot comparison is the closest like-for-like on the inbound half.

What Autocloz does not do for the inbound half

Being specific here matters more than being flattering.

It is not a content management system or a search-optimisation platform. It does not host your blog, manage your site's metadata, or track keyword rankings. Growzo, a separate open-source and self-hosted product, covers that ground; Autocloz does not, and no amount of configuration makes it.

It does not do paid-media buying. Ad lead capture from Meta, Google and LinkedIn lands leads into the workspace, but the campaign management, creative and bidding all stay in the ad platform.

Its lead scoring does not update itself. The rule engine supports seven rule types — title match, industry, company size bucket, country, tag, recent engagement and email verification — and clamps the result to 0–100, but the score is written by an explicit recalculation, not by a background sweep. If you build a qualification bar on the score, something in your process has to trigger the recalculation.

Its workflow triggers do not fire on every lead-creation path. A lead created through the API fires the lead.created trigger; a lead created by the contact-form capture path is written into its destination Leads List directly, which starts the attached campaign but does not run a lead.created workflow. If you are routing inbound with a workflow, attach the rule to the list rather than to creation, and test it with a real submission rather than assuming.

And it does not tell you which motion is working. It records what happened; the accounting judgement about which spend produced which opportunity is yours, and customer acquisition cost explains why that judgement moves the number more than any tool does.