Leadway — Campaign Brief

The read

Six things, in order of how much damage each one does.

He is optimising for free trials, not calls, and that is the right call. It also changes the entire shape of the campaign, so it goes first. At $29 a month a sales call costs more than the customer is worth, and inserting one reads as a company that does not trust its own product. So there is no calendar link anywhere in this sequence, no twenty-minute ask, and no discovery. The next step after every message is the link. The campaign gets judged on trials started, not on replies received, and the copy below is built backwards from that.
Two numbers decide whether this works and neither of them is public. Trial-to-paid conversion, and how long a paying account survives. At $29 the campaign runs on cost per activated trial, and without those two figures nobody, him included, can say whether a hundred trials was a good month or an expensive one. This is not a nitpick. It is the difference between scaling this in October and quietly shelving it.
There are three different ad-spend figures on one public profile. The headline says 7,000,000. The English About says "over 4 MIL USD in ad spend". The Russian version of the same About says 2,000,000. Anyone who switches the profile language, and a fair number of his target market can, watches the story change. Every message sends the reader to that profile.
Every certification listed has expired. Google Search Ads ran to Feb 2022. Google Analytics to Mar 2021. HubSpot Content Marketing to Oct 2021. He is selling data about Google Ads advertisers, so the expired Google credential is the one a sceptical reader checks, and it takes about ten seconds.
He publishes constantly and reaches almost nobody. Roughly fifty posts across July and August, most landing between zero and five reactions. Read that the right way round. He already believes in the channel, already does the work, and gets almost nothing back. That makes outbound the obvious fix rather than a new thing to be convinced of, and it means the campaign will not die of neglect.
The site sells the tool. The market buys the list. "Better Leads, More Business, Less Effort" describes a category with fifty products in it. What he actually has is 7,784 companies with live ads and 3,549 named contacts, filterable by category and country. For an agency, that is not a tool, it is a prospect list of people already spending money on the exact thing the agency sells. That distinction is the campaign.

The thing nobody is using: the product and the personalisation engine are the same object. Every message in this campaign is personalised by the thing being sold, which means the research cost per lead is a database query rather than twelve minutes of manual work. Nobody else selling into agencies this month can say that, and it is why this campaign can run at volume when the unit economics demand volume.

Why this offer

The campaign asks for a signup, not a meeting. That was settled by his own funnel before any offer was designed, and it rules out an entire family of otherwise good ideas. Three directions were then argued against each other on reply rate, effort to say yes, whether the input is free at volume, and whether the artefact survives being given away.

	Direction	Verdict
1	The category slice Three named advertisers in their niche, then the rest in the tool	Chosen. Useful standing alone, so it does not read as a teaser. Costs one query to produce. Asserts nothing about the reader. And it makes the trial the obvious next click rather than a favour they are asked for, because the three names are the proof that the other two hundred exist.
2	The whole list, sent Full category export in the message, no signup	Highest reply rate of the three and it eats the business. Once the list is in their inbox the tool is redundant, they have no reason to open it, and the thing he sells has been given away to someone who never became a user. Kept as the reply to a hard no on signing up, where a partial export is a cheaper concession than a discount.
3	The tool walkthrough Twenty minutes, screen shared, guided setup	Killed on arithmetic before it was killed on anything else. A call costs more than a $29 customer is worth, and it asks the buyer to pay in time for something they expected to click. It also contradicts the funnel he already runs. Kept for one case only: an agency asking about bulk or team access, where the price stops being $29 and a call starts paying for itself.

The slice is given away, genuinely, and the product holds the rest. Three real companies, named in the message, that the reader can verify in thirty seconds. Not a sample of a sample. If the three names are not immediately useful the trial looks like a paywall around something they already needed, and the message converts worse than sending nothing at all.

PULLEvery live advertiser in their niche, from his own database.

NAMEThree of them, in the message, by name.

OPENThe rest are in the tool. One click, no call.

The honest trade-off, and it runs the opposite way to a meeting campaign. Expect a much higher reply and activation rate, and far less control over what happens after the click. Somebody will open the trial, look once and never return, and no message can prevent that. Which is exactly why the instrumentation gate below is not optional paperwork.

The messages

Blank connection request. Sends from Andrei's personal profile, never a company page. Register is plain and slightly technical, because the reader is an agency operator. No em dashes, no emojis. No calendar link appears anywhere in the sequence, including in replies. Nothing is claimed about the reader's own accounts, clients or numbers at any point.

M1 · the three names50 frame · ~53 filled

Hi {{first_name}}, I track which companies are running live Google and Meta ads, by category and country. Right now there are {{category_count}} in {{category}}. Three of them: {{example_1}}, {{example_2}}, {{example_3}}. All three have ads running today. The rest are in a tool I built. Want me to send the link?

Three verifiable names is the entire mechanism. The reader can check one in thirty seconds, and once one checks out the count becomes credible without being claimed. Note what the message does not do: it makes no claim about their agency, their clients or their pipeline, so no lead needs verifying twice before it can send. "Want me to send the link" is the whole ask, and it is deliberately smaller than the thing being offered.

M2 · the second thing56 frame · ~56 filled

{{first_name}}, one more thing, and it's not from the ad data. From 1 September Google started auto-upgrading Search campaigns on automatically created assets or campaign-level broad match over to AI Max, rolling through the month. Accounts that launched recently get hit hardest, and those are the newest names on the list. Happy to send it over.

"It's not from the ad data" is doing provenance work. The list came from the ad libraries, this did not, and saying so stops the second message quietly borrowing the first one's authority. The migration is public, dated and live right now, which makes the re-ask new information rather than a nudge. It also does the real job of connecting a piece of industry news back to the list, so the list looks timely rather than static. Nothing is claimed about the reader's own campaign settings, because those are not public.

M3 · the close25 words

{{first_name}}, it's 7,784 companies with live ads and 3,549 named contacts, filterable by category and country. Say the word and I'll send the link over.

Both numbers are lifted straight off his own site, so they are safe to send today. The spend figures are not, because there are three of them and they disagree, which is why no message in this sequence carries a dollar amount until the third gate clears. There is also no open slot in this line. An open slot in a credibility claim is precisely where an invented statistic appears at fill time, and the close is the message least likely to get re-read before sending.

The recon

Six checks per lead, roughly four minutes, and steps one to three are a query rather than research. That is the whole reason this campaign can run at the volume $29 demands: the per-lead personalisation cost that would sink anyone else is a filter he already built.

1 DO THEY SELL TO ADVERTISERS? Agency, consultancy or in-house team whose buyers spend on ads. Their own site says it in the first screen or it does not. Sells to consumers, or sells something unrelated = DROP. 2 NAME THE NICHE THE WAY THEY DO Take {{category}} off their own homepage, not off a taxonomy. "Ecommerce paid social" not "marketing". If their site cannot name it in five words, DROP. An unnamed niche has no slice. 3 CUT THE SLICE Filter the database to that category and their market. {{category_count}} = live advertisers in it. Fewer than 25 = DROP. A thin slice makes the tool look empty, and the count is doing more work in M1 than the names are. 4 PICK AND VERIFY THREE NAMES {{example_1}} {{example_2}} {{example_3}}. Recognisable over impressive. Confirm all three have ads live in the Transparency Center ON THE DAY OF SENDING. Any one of the three dark = swap it, do not send. 5 SUPPRESS Existing users, expired trials, anyone already in the sequence. Pull that list before sourcing, not after. 6 LOG IT Category, market, count, the three names, date verified, cell. One row per lead so nothing is reused, contradicted or sent twice.

Step four is the campaign, not admin. The three names are the only claim in M1 that can be wrong, and a dead advertiser in the message is worse than no message, because it disproves the product in the same sentence that introduces it.

Targeting

The buyer is whoever owns finding new clients at a business that sells to advertisers. That is a narrower group than "agencies" and a much better one, because the list is literally their prospect list.

TITLES Founder · Co-Founder · Managing Director · Agency Owner · Head of New Business · Business Development Director · Head of Growth · Partner · Head of Sales at agencies and martech vendors EXCLUDE Delivery-side roles with no pipeline responsibility: Account Manager, Media Buyer, PPC Specialist, Designer, Developer, Project Manager, Analyst. Also anyone whose buyers do not spend on ads. SIZE 2-50 employees. Above 50 there is usually a data budget, an incumbent vendor and a procurement step, and a $29 tool does not survive procurement. Below 2 there is often no pipeline function at all and the buyer is the operator. SECTORS Paid media and performance agencies · full-service digital agencies · SEO and CRO shops selling upward into paid · freelance media buyers running a book · martech and adtech vendors selling to advertisers · white-label PPC providers TENURE Any. Ownership is the qualifier here, not tenure. GEOGRAPHY United States first, then United Kingdom. His own dataset is 68.5% US and 18.6% UK, so those are the two markets where a slice is dense enough to impress. Eastern Europe is 10.6% and is his home market, so it runs as a third cell with a different, warmer opener. One market per batch. The Transparency Center serves by country, and a session in the wrong country verifies the wrong ads at step four.

Signals - what makes this timed rather than demographic

Captured, primary. A hiring post for new business, sales or SDR at an agency. Someone spending money on pipeline this month is someone who will click a prospect list this week.
Captured, alternate. An agency under two years old. No client book yet, and finding advertisers is the whole job.
Created, the second key. A post about prospecting, lead generation, cold outreach or where their clients come from, inside ninety days. It is the reader telling you the problem in their own words.
The stack rule, hard. Being an agency is not a trigger. There are hundreds of thousands of them and on its own the filter selects for nothing, which turns the campaign into a list blast with a good artefact attached. Agency plus a pipeline signal is what sends.
Anti-signal, drops the lead. An existing data vendor named on their site, in a job post, or in their stack. Displacing a paid tool needs a comparison and there is not one yet.
Anti-signal, drops the lead. Over 50 staff. Procurement kills a $29 line item slower and more expensively than a no would.
Anti-signal, drops the lead. Their category returns under 25 live advertisers. The slice is the offer, and a thin slice argues against the product.

Volume: this campaign is judged on trials, so it wants scale. 25 connection requests a day, 800 to 1,000 leads in the first cycle, and the honest expectation is a low single-digit percentage starting a trial. That is a workable number at $29 and a terrible one at $2,900, which is the whole reason the ask is a link.

Before it sends

Eight things have to be true. The first two are the ones that decide whether this campaign can be judged at all.

Confirm the exact trial mechanics, in writing, before a word is writtenDoes it take a card. How many days. What is gated during the trial and what is not. "No card needed" is the single highest-converting phrase available to this campaign and it is not in the copy above, because it has not been confirmed to be true. If it is true, it goes in M1 and the reply rate moves. If it is not, no message may imply it.

Instrument trial-to-paid conversion and retention before launch. This one is silent and it is fatal.At $29 the campaign is judged on cost per activated trial, not on replies. Without conversion and retention wired up, a hundred trials in month one is an unreadable number and the decision to scale or stop gets made on a feeling. Tag every campaign signup so the cohort can be followed separately from organic. This is the difference between a pilot and an expense.

Reconcile the three ad-spend figures7,000,000 in the headline, "over 4 MIL USD" in the English About, 2,000,000 in the Russian About. Pick the true one and correct the other two before any connection request goes out. Every message sends the reader to that profile, and the first thing a marketer does with a number is look for the other one.

Renew or remove the expired certificationsGoogle Search Ads expired Feb 2022, Google Analytics Mar 2021, HubSpot Oct 2021. Selling data about Google Ads advertisers with an expired Google Ads certificate on the profile is worse than having none, because it dates the last time the credential mattered enough to renew.

The three names are verified live on the day of sendingNot the day the list was built. An advertiser who has paused since the export turns M1 into a disproof of the product, in the same sentence that introduces it. This is the only claim in the sequence that can be wrong, so it gets checked last, not first.

Say what "verified" means on the site, and how fresh the data isThe homepage claims verified domains and verified contacts. The first question a buyer with a data budget asks is verified how, and refreshed how often. Agree the honest answer now, because it will arrive in a reply within the first week and an improvised answer will be wrong.

No dollar figure in any message until the third gate clearsNot the 7M, not the 4M, not the $20M in client revenue. Until the profile agrees with itself the campaign runs on the two site figures, 7,784 and 3,549, which are published, current and checkable. This is a copy instruction, not a preference.

Sending capacity, and whether Sales Navigator is neededHis account is not Premium, so there is no Sales Navigator and no InMail. At 25 connection requests a day inside free limits the volume works, but 800 to 1,000 leads have to be assembled somewhere other than Sales Nav. Decide before sourcing starts, because it changes how the list gets built and what it costs.

What I'd fix on his profile

Every message sends people to this profile, and right now it argues with itself. Three moves, all prerequisites rather than improvements.

One number, everywhere, in both languages. Whichever figure is true. The current spread is the only thing on the profile a reader can catch him on, and it sits in the headline.
Lead with the list, not the career. The headline sells a performance marketer. The campaign sells a database of advertisers. A reader who clicks through from M1 should land on the thing they were just offered, not on a CV.
Put the three-name proof on the profile too. A pinned post showing a real category slice does the same job as M1 for anyone who arrives without one, and it is the only piece of proof he currently has that costs nothing to produce.

Fastest route to real proof: two named agencies willing to say the list found them a client. Not a testimonial about the interface. At $29 nobody needs convincing the tool is affordable, they need convincing the data is real, and one named outcome does that better than any feature list.