Agency Lead Generation When Your Buyers Run Stores

Agency lead generation reads two ways: lead generation for an agency, and agencies that sell lead generation as a service. This post is the first one. It is about how an agency selling to ecommerce brands fills its own pipeline, and specifically about how it decides which stores to write to this week.

The disclosure first. We build Keaz Signals, an ecommerce leads database with live buying signals, so we sell into exactly the problem this post describes. We are not a neutral party and there is no point pretending otherwise. What we can do is be precise about where the approach works and where it does not, which is the part most articles on this term skip.

The audience here is narrow on purpose: performance and Meta agencies, retention and Klaviyo specialists, shop developers, creative and video studios, and software companies selling into ecommerce. Our own site states the qualifier plainly — the clearest sign this fits is that you have paid sales capacity that is not fully booked. If referrals fill your calendar, outbound is a solution to a problem you do not have yet.

One more boundary. Nothing below is about how to run a store. It is about how to sell to the people who do.

Almost no agency is short of stores to contact. As of August 2026 we watch about 4.1M Shopify and WooCommerce stores across the EU and North America — 1.7M in Europe, 2.4M in North America. Any competent directory or scraper will hand you tens of thousands of them. Volume is not the scarce thing.

What is scarce is a reason to write to a particular store in a particular week. The failure mode of agency lead generation is not an empty list, it is a worn-out one. Everyone in your category is drawing from the same well, so the founder you are writing to has already had three near-identical emails from three near-identical agencies, and yours arrives fourth with the same opener.

That is a targeting problem wearing a volume problem's clothes. The instinctive fix — send more, buy a bigger list, add another sending domain — raises cost on both sides of the ledger. You pay for more addresses, you burn more sender reputation, and you get a lower yield per email because nothing about the message has become more relevant to the person reading it.

There is also a decay problem nobody puts on the pricing page. A list is accurate on the day it is exported and stops being accurate from that moment, and nothing in the file tells you when it stopped. Firmographic attributes — platform, country, category, employee count — are stable, which sounds like a virtue until you notice that a fact which never changes can never tell you that now is the moment.

So the useful question is not how to get more stores. It is which forty of the ones you already have are in a buying window this week.

A store event is something the business did on a dated day: new Meta ads going live, the active ad count rising, a product launch, an email marketing tool installed, social growth on Instagram or TikTok, a change in newsletter rhythm, or a storefront rebuild. Each one carries a date, and the date is what makes it usable.

Read down that list as an agency and most rows are already addressed to someone specific:

• New Meta ads running, or ad count rising — a performance or Meta agency's clearest opening. Budget is moving right now and someone is watching the numbers.
• An email marketing tool installed (Klaviyo, Brevo and the rest) — retention work has just been funded and is usually unstaffed for the first month.
• A product launch — creative, video and launch-campaign work, on a deadline the store set itself.
• Store and site changes — a storefront rebuild is a shop developer's brief already in progress, whether or not it is going well.
• Social growth and posting cadence — content, community and UGC teams, where a rising follower count is evidence the channel is being taken seriously.
• Newsletter activity — lifecycle and copy work, and a reasonably honest proxy for whether anyone owns the channel internally.

The full set is on the buying signal catalogue, with what each one means and how it is detected. A worked example for one of them — how a Meta ads agency turns two dated events into a week's shortlist — is in how Meta ads agencies find stores worth pitching.

The dating is not a detail. A signal with a date attached tells you two things — that it is worth acting on, and when it stops being worth acting on. We do not present a three-week-old event as an opportunity, because it is not one. An attribute with no date can only ever tell you the first half of that, which is why a filter on platform and country produces a list and a filter on last week's events produces a shortlist.

Freshness is the constraint that makes or breaks this. Stores are re-scraped weekly, so what you see is days old rather than weeks. That number matters more than coverage does: a signal you find a month late is a firmographic attribute with extra steps.

Four steps, and none of them is "buy a bigger list". Build one segment that matches the work you actually want. Let the events accumulate against it. Each week, write only to the stores where an event fired in the last seven days. Send from your own infrastructure so the replies come back to a place you control.

1. Segment once, not weekly. Country, platform, vertical, follower counts, ad activity and the rest, with a live audience estimate as you narrow it. The point of doing this once is that the segment describes the work you want, and the work you want does not change every Monday.
2. Watch it instead of re-exporting it. This is the actual difference between a list workflow and a signal workflow. The list workflow asks "who is in my market" and gets the same answer every time. The signal workflow asks "who moved", and gets a different, much shorter answer each week.
3. Write from the event, not from the firmographic. "You run a Shopify store in Germany" is not a reason to email anyone. "You put four new ads live last Tuesday" is. Our sales agent writes the per-store copy from the signal plus your own knowledge base, but the principle stands whether a machine or a junior does the writing.
4. Send through your own sending stack. Campaigns are created in your own Instantly workspace; status and replies sync back. Your domains, your warmup, your reputation — which also means the sending decisions stay yours rather than being pooled with strangers.

The step most agencies skip is the second one, because re-exporting feels like progress and waiting does not. It is worth being blunt about the arithmetic: a segment of forty thousand stores where two hundred did something this week gives you a Monday shortlist of two hundred, and two hundred well-timed emails is a smaller, cheaper and more answerable week's work than four thousand untimed ones.

Two adjacent pieces go deeper on the sending half of this. What a sales engagement platform does covers why the tool that runs the sequence is usually silent on who belongs in it, and cold email subject lines driven by a buying signal takes the same argument down to the first line of the email. If you are weighing whether to automate the writing step at all, what an AI SDR actually automates is the honest version of that question.

Plainly, and without hedging. If the businesses you sell to are not Shopify or WooCommerce stores in the EU or North America, none of this reaches them. If you need the leads out as a file, we do not export — leads move into campaigns and stay there. For some teams that is a dealbreaker, and it should be.

There are no LinkedIn signals and no LinkedIn sending here, and no plans to add them. If the people who buy from you spend their working day posting on LinkedIn, buy the tool that watches LinkedIn. A store-level signal will not see them, and we would rather say that than sell you a partial answer.

Registry and firmographic triggers, and hiring signals, are marked Planned on our own site. Planned means not shipped. If your qualification depends on knowing that a store just registered a new entity or opened a role, that is not something you can do here today, and no amount of adjacent capability changes it.

Access is capped per market, which is a real constraint on you rather than a scarcity device. If your market is full, you wait — and if you need to start outbound on Monday, that is a reason to look elsewhere this week.

One thing that is not a limitation of ours but of the category: signal-led outbound does not fix an offer nobody wants. Timing multiplies whatever the message is already worth. If the underlying pitch is not landing, a fresher list changes the arithmetic and not the outcome. That is the least commercially convenient sentence in this post and it is also the truest one.

Where we sit against the other shapes of tool — store directories, contact databases, general sales-intelligence platforms — is laid out on our comparison pages, including the rows where the other product wins. The structural version of the same argument is in what a sales intelligence platform knows, and what it cannot.

One number, ours, dated. Running signal-led outreach on our own pipeline, as of August 2026, contacting a store in the week its event happened moved reply rate from about 3.9% to about 8.5%, with no change to the message. The only variable was timing.

Now the limits of that number, which matter more than the number does. It is one sender, one market, one offer, over a window we chose to look at. It is not a controlled trial and we are not going to dress it up as one. It moves, so it carries a date every time we use it, and a figure from August 2026 is not a claim about next quarter.

We are also not going to promise you a reply rate. Nobody honestly can. A vendor quoting a reply rate is quoting their best case as your baseline, and the difference between those two things is your market, your offer, your domain history and your list hygiene — none of which they have seen.

What survives all those caveats is the structural part: the message did not change. If the same email performs differently depending only on when it lands, timing is a lever sitting unused in your process, whatever its size turns out to be in your market. That is a claim you can test on your own sending in a fortnight, which is a better use of your scepticism than arguing with our number.

One more of our own figures, because it decides whether any of this is reachable at all: contact coverage across the database was 84% as of August 2026, including founder addresses that are not listed publicly. Also a moving number, also dated. What we will not tell you is a per-signal precision or a bounce rate, because we have not measured those in a way that would survive being quoted.

Signing up gives you 1,000 leads free, spent inside segments you build yourself. That is enough to test the premise on your own market: pick one signal, take forty stores where it fired in the last seven days, write to them from the event, and compare that week against whatever you sent last month.

Run it as a comparison rather than a launch. Keep the offer, the sender and the sequence identical to what you already do, and change only which stores receive it and when. If you change three things at once you will learn nothing, which is how most outbound experiments end.

Forty is a deliberately small number. It is small enough to write properly, small enough that you will notice the replies individually rather than as a percentage, and small enough that a bad week costs you an afternoon. Scale is the reward for a result, not the way to get one.

If it does not move, you have learned something cheap and specific about your market, which is more than a bigger list would have told you. If it does, the question becomes which signal did it, and that is a much better problem to have on a Monday than "where do we find more stores".

The segment builder and the coverage figures are on the ecommerce leads database page, and what a seat costs once the free credits run out is on the pricing page. Access opens per market in order of signup, so the honest answer to "can I start today" is: depends on your market.

Every figure in this post is our own, taken from our own pages and our own sending. Nothing here is quoted from a third party.

• Keaz Signals ecommerce leads database — coverage (4.1M stores; 1.7M Europe, 2.4M North America), contact coverage 84%, and the 1,000 free leads on signup. Retrieved 3 September 2026; page states "Updated August 22, 2026".
• The buying signal catalogue — the signal types listed above and how each is detected. Retrieved 3 September 2026.
• Keaz Signals sales agent — per-store copy from the signal, and the Instantly sending path. Retrieved 3 September 2026.
• Our own outreach, as of August 2026 — reply rate of about 3.9% rising to about 8.5% when contact fell in the week of the event, message unchanged. One sender, one market, one offer; not a controlled trial.
• Keaz Signals compared honestly — the export, LinkedIn, platform and market limits named above. Retrieved 3 September 2026.