---
title: "How we closed seven automated facilities worth $8,381,800 for a warehouse automation integrator in ten months. | Stone Haven Capital Group"
description: "An average of $1,197,400 per facility. Ten months. A goods-to-person robotics and warehouse software integrator. No new sales hires, no systems-integrator channel, no waiting for a distribution centre director to publish a tender."
canonical: "https://stonehaven.capital/showcase/warehouse-automation-integrator"
last-updated: "2026-08-22"
---

> An average of $1,197,400 per facility. Ten months. A goods-to-person robotics and warehouse software integrator. No new sales hires, no systems-integrator channel, no waiting for a distribution centre director to publish a tender.

Stone Haven Capital Group

Schedule a Call

Case study · Warehouse Automation Integrator

# How we closed seven automated facilities worth $8,381,800 for a warehouse automation integrator in ten months.

An average of $1,197,400 per facility. Ten months. A goods-to-person robotics and warehouse software integrator. No new sales hires, no systems-integrator channel, no waiting for a distribution centre director to publish a tender.

- $8,381,800 generated in 10 months, from 7 automated facilities at an average of $1,197,400.

- 288,078 emails to 48,013 companies, producing 624 replies at 1.3%.

- 125 interested conversations became 35 held meetings and 7 signed facilities.

- What did most of the work: we kept every cold send off the client's own domain and built a separate estate for it.

- What went wrong: the first cadence was built for a sales cycle a quarter of the length of this one.

01

## Did they get a good result?

Yes. Here it is plainly. In the twelve months before we started the company had signed $2,793,933 of facility work, all of it arriving through two 3PL relationships and the occasional referral from an equipment vendor. Ten months later seven facilities had been signed for $8,381,800, an average of $1,197,400 each, with 35 qualified opportunities worth $31,150,000 still open behind them. Three times the baseline, off seven signatures.

What happenedThe number

Facilities signed in ten months7

Total contract value signed$8,381,800

Average value per signed facility$1,197,400

Qualified opportunities behind them35, worth $31,150,000

Interested-stage conversations125, worth $71,843,000

Emails sent across the programme288,078

Buyers contacted48,013

Replies received624, a 1.3% reply rate

Positive replies125, 20.0% of all replies

Bounces across the estate2,305, a 0.8% bounce rate

Ten months of sending against a buyer pool that fits in one building.

02

## Did you get them in front of the right people?

This is the only question the send count cannot answer, and on this programme it is close to the only question at all. Seven signatures across ten months means each one is fourteen percent of the result. There is no volume anywhere in this shape to absorb a targeting mistake.

35 opportunities cleared qualification carrying $31,150,000 between them, an average of $890,000 each at that stage. Seven of those signed, for $8,381,800, at an average facility of $1,197,400. Behind them sat 125 interested-stage conversations worth $71,843,000, roughly $574,700 apiece.

Targeting shows up in what survives the funnel. 624 replies produced 125 positive ones, 35 of those reached qualification and 7 signed. A fifth of every reply was genuinely interested, and more than a quarter of those interested conversations reached a qualification bar set at a real facility with a real budget line. Those two ratios do not happen to a list that was bought.

The seven signed facilities and the qualified work still sitting behind them.

03

## What worked, and why

We kept every cold send off the client's own domain and built a separate estate for it.

An integrator whose software runs inside a customer's distribution centre cannot afford a deliverability incident on the domain that carries its support tickets and its go-live coordination. So none of the outbound touched it. 44 dedicated domains carried 132 inboxes, each one authenticated on SPF, DKIM, DMARC and MX before it sent anything. Warmup ran continuously on all 132 for the life of the engagement, and only a portion of the estate carried campaign volume at any point, with the rest ageing in reserve.

That estate carried 288,078 sends and finished at a 0.8% bounce rate, 2,305 bounces in total. Inbox placement held at 94.0%. Each inbox was capped at 20 sends a day, which is what keeps a domain alive across ten months rather than three. The infrastructure logged four alerts and no delivery errors across the whole run.

The estate that carried the programme. None of it on the domain the client's customers know.

The reason this matters more here than usual: the buyer pool is small enough that the same operations director will hear from this client again in two years. Burning a domain does not cost you a campaign in this market. It costs you the only list you have, permanently.

### We built the list around facilities rather than companies, then cut it in half.

The person who signs an automation project is a VP of operations, a distribution centre director or a VP of engineering, at a third-party logistics provider or a retailer running its own network. The first cut was that profile across North America. The second cut was the one that mattered: it kept only organisations whose facility footprint, throughput profile and stated network plans made a goods-to-person build physically plausible in the next eighteen months.

Out went single-site operators, anyone whose ceiling height or lease term ruled out the hardware, pure-play parcel carriers, and every company already three years into a competing platform. That removed roughly two thirds of the starting pool before a word was written.

What that buys: a distribution network is a fixed asset. You cannot sell a robotics build into a building that will not take one, and you cannot find that out on a call without wasting an engineering pre-sales day. The list has to carry that judgement, because nothing downstream of it can.

### We wrote an opener that qualified the facility and asked for nothing else.

The opener greets the person, states one capability with the credential that makes it believable, names the kind of operation it applies to, and asks a single question about their own network. It closes with a full signature and a title, because a seven-figure capital purchase does not get answered by someone who signs with a first name. One claim, one proof, one question.

Everything left the estate as plain text, varied line by line, so no two operations directors inside the same retail group received identical wording. Two versions of the opener ran throughout, differing only in which consequence of the build appeared first.

The opener that carried the whole programme, two variations and nothing else.

The gap between those two lines is the honest measure of an opener on a programme like this. A question that is easy to answer and hard to keep produces a calendar full of accepted invitations and an empty room. What we watch weekly is what percentage of created meetings survive to a held conversation, because that number moves before anything in the pipeline does, and it is the earliest signal that an angle is charming people rather than qualifying them.

### We ran six touches on widening gaps and let the sequence live for seven months.

Six steps with gaps of four, seven, ten, fourteen and twenty-one days, stretching each contact's sequence across most of a quarter, running against a live list for seven of the ten months. Two opener variations, both A/B tested. Every touch after the first carried its own reason to exist rather than restating the one before it.

Why the gaps widen: the first two touches are testing whether the message landed at all, and speed is useful there. Everything after that is waiting for a capital planning cycle to turn over. A twenty-one day gap in month four reads as a person who is still around. A four day gap in month four reads as software.

### We judged angles on conversations rather than counts, and accepted that took months.

624 replies across ten months is a handful a week. At that rate you cannot separate two angles statistically inside a fortnight, and pretending otherwise is how a programme kills its best message on noise. So each angle accumulated replies until the difference between them was wide enough to act on, and the weekly review read the actual conversations instead of a leaderboard.

04

## The three angles we test, in every market

Everything above is where the programme ended up. It did not begin there. At the start nobody knew which of three messages this market would answer, and no amount of reasoning gets you to that answer in advance. Three is a deliberate number here. A market of a few hundred genuine buyers gives you one approach per contact and no second chances, so every additional angle thins the evidence base underneath all of them. Five angles against 624 replies would have told us nothing about any of them.

05

## 1. The teardown

Take one thing about their operation that you can see from outside and show the work. A read on their throughput profile, a picture of where the labour hours are going, a number they have never had put in front of them. It earns the reply by proving you looked before you wrote.

We mapped what a peak shift looks like across the three buildings in your northeast cluster and where the pick hours actually go. Want me to send the read over?

06

## 2. Social proof

Put a comparable operation in front of them with a real outcome attached and let them draw the parallel themselves. In a market where every buyer already knows every other buyer, this one carries further than it does anywhere else.

A 3PL running a similar case-pick profile went live on goods-to-person last year and took a third out of their peak labour plan. Worth the short version of how it was phased?

07

## 3. Value first

Open with something useful whether or not they ever answer. A benchmark, a build cost range, a lesson from a deployment that went badly. Usefulness first, interest second, and no obligation attached to either.

We put together what the first year actually costs on a goods-to-person build, including the integration work nobody quotes for. Happy to send it across.

08

## How we use them

All three launch at once against a divided list, and then nothing happens for a while. That is expected. With this reply volume the split takes months to resolve rather than weeks, and the discipline is refusing to call it early. Once the winner is clear the other two come off and their share of the list goes to the survivor, which is the only way a pool this small gets used well.

09

## What did not work, and what we did about it

Three of them. You should hear these from us rather than work them out in month four of your own programme.

The first cadence was built for a sales cycle a quarter of the length of this one.

We opened with a short sequence on tight gaps, which is the right shape for a market that buys in weeks. This one buys in quarters. Contacts finished the sequence and went quiet while the buyer was still three months away from the meeting where next year's capital gets allocated. We were not being rejected. We were arriving, finishing and leaving before the conversation that decides anything had started.

The starting pool. Every operations and engineering lead at a logistics operator of roughly the right size.

What we changed: the sequence was rebuilt to six touches on widening gaps running seven months, so a contact entered in week two is still in it in month six. The list was rebuilt underneath it at the same time, from 165,561 contacts down to 55,187 once facility profile and platform exclusions went on. Fewer people, each of them reachable across an entire budget cycle rather than a fortnight of it.

The same market after facility profile, footprint and platform exclusions.

### A named-account list was worked by two people at the same time.

Ownership of the strategic accounts was never written down, so two people worked the same names in parallel. Two prospects received the same email twice on the same day, from two senders, at two different domains. In a market where the buyers know each other, that is not a duplicate send. It is a story someone tells at a conference.

What we changed: named accounts moved to a single owner with hard suppression across the whole estate, so a contact touched by one sender is invisible to every other. Nothing about the messaging changed. The failure was administrative, and it was ours.

### We attached case study PDFs to a follow-up and the estate noticed within days.

A mid-sequence touch went out carrying two PDF case studies, on the reasoning that a seven-figure buyer wants evidence. Bounce rate on the affected domains climbed off its baseline inside four days and inbox placement on the same domains fell below the rest of the estate. Attachments from an unknown sender are treated exactly the way you would expect them to be treated. The evidence was fine. Delivering it that way was not.

What we changed: the attachments came off, the material moved behind a link sent only after a reply, and the affected domains sat out of rotation while their placement recovered. Across the full programme the estate still finished at 0.8% bounce and 94.0% placement, which is what a fast correction buys you.

A programme with seven signatures in it has nowhere to hide a mistake, and a document showing none has had this section deleted. These are the corrections that turn up again on your engagement, which is the only reason they are worth your time.

10

## Why this works for warehouse automation specifically

The integrators who come to us describe the same situation. The technology works, the reference sites are real, and every new facility still arrives through a 3PL that already knows them or an equipment vendor passing something along. That builds a genuine business with a completely unforecastable one attached to it, and it never reaches the retailer three states over who is quietly planning two new buildings.

This market suits the channel for three specific reasons:

- The buyer pool is small enough to name. A few hundred organisations in North America can actually take a goods-to-person build. That is a list you can enumerate, exclude down and then work properly, which is a luxury almost no market offers.

- One facility funds years of the channel. At an average signed facility of $1,197,400, the arithmetic stops being a marketing question after the first signature. That ratio is what makes a seven-deal programme rational instead of reckless.

- The trigger is public and early. Network expansion, new lease announcements, throughput commitments and engineering hiring all appear well before a capital request is written. The timing problem solves itself if you are reading the right signals.

11

## Why would this work for your business?

It may well not, and the cheapest moment to discover that is before a single domain is bought. Work down the table.

This works ifThis does not work if

A single signed project is worth six or seven figuresYour average sale is a few thousand dollars and needs volume to matter

Your buyers can be enumerated: hundreds of real organisations, not hundreds of thousandsYour market is genuinely mass and anyone with a budget could buy

The purchase attaches to a capital cycle you can wait outYour revenue plan needs the channel to produce inside sixty days

You have installed references you are allowed to nameEvery deployment is confidential and nothing can be pointed at

A senior person can hold a technical conversation within a week of a replyReplies would sit unanswered until someone frees up

If the left column describes you, what transfers is the method rather than anything unusual about this client. An integrator with real installed references, a capital purchase attached to a fixed asset, and no route to an operator who has never heard of them. All of it came out of process, and the process does not change for you.

12

## One more thing worth understanding

A ten month engagement in a market like this books almost nothing in its first ninety days, and that is the design working rather than the design failing.

The reason is arithmetic. With a six touch sequence spread across widening gaps, a contact entered in week two does not reach their final touch until month four. The angle split needs months of replies before the difference between two of them means anything. The estate is warming into full volume across the same period. Day sixty finds a programme where almost every input is still mid-flight and almost no output has landed.

Judged at day sixty this one looked like a failure. A handful of replies, no signatures, and a cost line with nothing next to it. Every honest thing you could say about it at that point was discouraging. The first facility signed in month four and the seventh in month ten, and five of the seven came from contacts first emailed before day ninety.

So the thing to agree on before starting is what gets measured when. In the first quarter it is reply quality and whether the conversations are with people who own a facility decision. Pipeline is a month five question. Signatures are a month seven question. An engagement cancelled at day sixty for the reason that looks most obvious is one where the money was spent and the return was thrown away two months before it arrived.

13

## Before and after

Before After

Where new facilities came fromTwo 3PL relationships and occasional equipment vendor referrals 48,013 named buyers approached directly

Signed facility value in the period$2,793,933 $8,381,800, three times the baseline

Qualified work visible at any momentWhatever the referral network happened to be carrying 35 opportunities worth $31,150,000

Conversations behind the forecastEffectively none until a referral arrived 125 interested-stage conversations worth $71,843,000

Sending infrastructureThe company domain, used for everything 44 dedicated domains carrying 132 inboxes at 94.0% placement

14

## If you want to know whether your market has this in it

A quarter of an hour usually settles it. Tell us which operators you want inside, what a facility is worth when it signs, and how long one takes from first conversation to purchase order. We come back with how many of those organisations exist, what conversation volume is realistic against that number, and a direct answer on whether this channel suits you.

If the answer is no, we say so on that call. Neither of us gains from discovering it in month six of something that was never going to work.

## Want to know whether your market has this in it?

The first conversation is short. You tell us who your buyers are and what one is worth to you. We tell you how many we can actually reach, what the meeting volume looks like, and whether outbound is the right lever for you at all.

If we think it is not, we will say so.

Book a consultation call

→

## Read another one

[AI Implementation Consultancy](https://stonehaven.capital/showcase/ai-implementation-consultancy/)[How we closed $4,320,000 of AI deployment work for an implementation consultancy in six months.](https://stonehaven.capital/showcase/ai-implementation-consultancy/)[Aerospace & Defense Tier-One Supplier](https://stonehaven.capital/showcase/aerospace-defense-tier-one-supplier/)[How we built a $31M+ active program pipeline for a tier-one aerospace and defense supplier.](https://stonehaven.capital/showcase/aerospace-defense-tier-one-supplier/)[All case studies](https://stonehaven.capital/case-studies/)