Jacquard sells language optimization to large brands, and the customer list is the strongest evidence in its favour: Adidas, Accor, United Airlines, Costco, Sephora, Nestle, Tesco, Best Buy. That is not a pilot-stage roster.
Guardrails are the right thing to lead with
The objection enterprise brands raise about generated copy is almost never quality. It is control. A brand with a style guide, legal review and a tone that took a decade to establish cannot hand message writing to a system that might produce something plausible and wrong.
Jacquard’s framing centres on generation inside defined brand constraints, which addresses that objection directly rather than arguing around it. For a buyer in that position it is the difference between a product they can deploy and one that dies in review.
Multi-language coverage across 27 or more languages is the other concrete differentiator. Tone consistency across markets is genuinely difficult, normally solved by a patchwork of local agencies with variable results, and doing it inside one system with the same optimization loop applied to each market is a real capability rather than a repackaged one.
The case study framing deserves scrutiny
Currys is the headline: a 42 percent uplift in opens, 93 percent on clicks, 102 percent on revenue.
The open figure should not be leading that sentence. Apple Mail Privacy Protection prefetches message content and conceals actual activity, which means reported opens include machine fetches. On a consumer list a substantial share of recipients sit behind it. An uplift in opens can therefore reflect who was sent to as much as whether anyone read.
The clicks and revenue numbers are the useful ones, and they are strong. Leading with the compromised metric is a choice, and in a category built on measuring response, choosing the least reliable measure as the headline is a reasonable thing to ask about on the first call.
The larger gap is that none of the three figures comes with a test design. A holdout group, a period, a sample size, and confirmation that nothing else changed at the same time are what separate a result from a coincidence. Enterprise programmes change several things per quarter, and before-and-after comparisons absorb all of them.
The volume question comes first
Language optimization detects differences between variants, and those differences are typically a few percentage points. Detecting a few percentage points reliably takes a large sample.
That makes send volume the qualifying criterion, ahead of budget or interest. A list of 50,000 will rarely separate a genuinely better subject line from a marginally different one within a single campaign, and running the same test across many campaigns introduces seasonality and audience drift.
Ask what volume per test the optimization requires to reach significance, and check it against your actual programme. If the answer is comfortably below your sends, the rest of the evaluation is worth doing. If it is above, no amount of product quality will help.
Against Persado
The two compete directly. Persado publishes more about its underlying language framework and testing approach, which makes the mechanism easier to interrogate before purchase. Jacquard’s advantages are the guardrail framing and the breadth of language deployment. For a single-market brand the choice is close; for a global one, the multi-language optimization is the concrete separator.
