Larta case study: 5.2% reply rate testing Poland

The client

Larta is a Latvian e-commerce retailer of farm and livestock supplies — milking equipment, electric fencing, feed and mineral supplements, incubators, agricultural machinery. In Latvia it sells both to farms directly and through retail. The question it brought to us was narrower than "can we grow": the owner wanted to know whether Poland was worth entering at all, before committing budget to it.

Who we actually targeted

Not farms. The target was shops — Polish retailers selling farm and livestock supplies, where the decision-maker is the owner rather than a procurement department. That distinction is the whole campaign: a shop owner can say yes to a new supplier in one conversation, and reads their own email.

The offer matched the audience. The first message did not pitch products or ask for a meeting — it asked whether they were open to working with a new supplier. For a shop owner, that is a question with an obvious yes-or-no answer, which is why it got answered.

The challenge

Poland is the largest market in Central and Eastern Europe, and two things make it hard for Latvian exporters: it runs in Polish, and it is stricter than most of the EU on cold email. On top of that, "shops selling farm supplies" is a category that no database defines cleanly — which turned out to be the campaign's real difficulty, and its most useful lesson.

How it ran

Preparation: one week — defining the segment, building and verifying the list, writing and approving the Polish copy.
Sending: 22 working days, weekdays, Warsaw business hours.
Language: native Polish, written rather than translated.
List: built by our own research process rather than bought — 950 identified, 847 delivered, 2 failed, 0 invalid.
Format: a fixed-scope Sales Pilot — one country, one segment, one language.

What the 44 replies actually were

This is the part that matters more than the percentage. The replies were not polite acknowledgements — they were requests for price lists, requests for catalogues, and questions about specific products. That is a buying conversation starting, not an open rate.

There were also zero unsubscribes across all 847 messages. In a market with strict electronic-marketing rules, a campaign that draws catalogue requests and no removal requests is telling you the targeting was right.

What did not work

We did not get the segment definition tight enough. "Shops selling farm and livestock supplies" is clear to a human and ambiguous to any filter — the category bleeds into pet retail, garden centres and general rural stores. Some of those are legitimate prospects. Some are not.

The message that made this concrete: one of our emails went to a shop that deals in spiders. A farm supplier has nothing to sell an exotic-pet retailer, and no wording would have rescued that send. It was a targeting error, not a copy error.

The honest read is that a segment this fuzzy needs a tighter disqualifier rule before sending, not better copy afterwards — the thing our targeting analyzer now checks first. On a pilot this is cheap to learn. At scale it would have been expensive.

Why it worked anyway

Three things carried it. Writing in Polish rather than sending translated English, which most Latvian exporters skip. Targeting a buyer who can decide alone — the shop owner — instead of an organisation. And asking one simple question instead of presenting a catalogue.

One pilot is one data point, and we present it as one. What it answered for Larta was the question they actually asked: yes, Polish shops will talk to a Latvian supplier, and here are 44 of them who want to see prices.