Case Study: Vegan Food Brand Grows to 5,000 Subscribers Using Personalisation
A UK vegan food brand grew to 5,000 subscribers by using personalised subscription boxes filtered by dietary preference and allergen data — reducing churn and increasing average order value.
The vegan food market is competitive, with dozens of brands competing for the same health-conscious, values-driven consumer. One UK vegan food brand differentiated on personalisation — using allergen and dietary preference data to create genuinely customised subscription boxes.
The Personalisation Strategy
Rather than a fixed weekly box, the brand offered subscribers a preference profile:
- Core dietary profile (vegan is baseline — all products are vegan)
- Allergen exclusions (gluten-free, nut-free, soy-free)
- Taste preferences (spicy / mild, adventurous / familiar)
- Calorie range preference
Each week, the system generated a personalised box recommendation from the 40+ SKU catalogue, filtered by the subscriber's allergen exclusions and taste preferences.
The Technical Implementation
Allergen Matrix provided the allergen data layer — each product tagged with full allergen declarations. SaltAI Subscriptions used this data to filter the subscription product pool for each subscriber, showing only genuinely compatible products.
The Growth Story
Launch month: 120 subscribers (founder network and Instagram followers) Month 6: 890 subscribers (organic content and word of mouth) Month 12: 2,400 subscribers (paid social, podcast appearances) Month 18: 5,100 subscribers
Monthly churn at 18 months: 7.2% — significantly below vegan subscription box average of 15–20%.
The brand attributed low churn primarily to personalisation — subscribers felt understood, not just served a generic product.
Build a personalised subscription food business with SaltAI's tools.
Why Allergen Data Became a Competitive Advantage
Most vegan subscription boxes treat allergen management as a compliance checkbox — something handled in the small print rather than built into the customer experience. This brand took the opposite approach, making allergen transparency a front-facing feature that appeared prominently in onboarding, product pages, and weekly box preview emails. Subscribers with coeliac disease or nut allergies reported that this was the first subscription box they had trusted enough to order without manually checking every product label themselves.
The allergen data layer also solved a significant operational problem: manual curation at scale. Before integrating Allergen Matrix, the team spent hours each week manually cross-referencing products against subscriber exclusions. After integration, the filtering happened automatically at the catalogue level, meaning the curation team could focus on discovery and product sourcing rather than safety auditing. This freed up roughly twelve hours of staff time per week by month six — a meaningful efficiency gain for a small team.
Beyond operations, allergen transparency became a genuine marketing asset. The brand featured real subscriber testimonials from customers with complex dietary requirements — people who described the box as genuinely life-changing for their weekly meal planning. These testimonials outperformed standard lifestyle content in paid social testing, achieving a 34% higher click-through rate in Meta ad campaigns. When personalisation solves a real problem rather than simply adding novelty, it generates advocacy that no paid strategy can replicate.
Building the Preference Profile: What Worked and What Didn't
The initial onboarding flow launched with eleven distinct preference fields — taste intensity, cuisine origin preferences, cooking time tolerance, ingredient familiarity, and several others alongside the core allergen and dietary data. Completion rates were poor: only 41% of new sign-ups finished the full onboarding sequence. A significant proportion abandoned mid-flow, either on the cuisine preference screen or the cooking time question. The team interpreted this as fatigue — eleven questions felt like a survey rather than a welcome experience.
The redesigned onboarding reduced the required fields to four: allergen exclusions, taste intensity, adventurousness, and calorie range preference. Everything else became optional, surfaced in account settings as an invitation to "refine your box" after the first delivery. Completion rates on the four-field version reached 89% within two weeks of the redesign. Crucially, the optional extended preferences were completed by 67% of subscribers within their first three boxes — people chose to engage when the pressure was removed from the initial sign-up moment.
The lesson applied beyond onboarding. Preference data collected passively — through box ratings, product skips, and repeat selections — proved more accurate than explicitly stated preferences. A subscriber who said they preferred familiar flavours but consistently rated adventurous products highly was served more adventurous options, and churn risk for that segment dropped noticeably. Building a feedback loop into the subscription experience, rather than relying entirely on the initial profile, was one of the most impactful product decisions the brand made in the first year.
Content Marketing as a Growth Channel
Organic content drove a disproportionate share of early subscriber growth, particularly during months three through nine before paid social budgets were meaningfully scaled. The brand published weekly recipe content built around products in the current catalogue — not generic vegan recipes, but specific preparations using the exact items subscribers were receiving that week. This created a sense of community between box recipients and gave potential subscribers a concrete reason to follow before committing to a purchase.
The content strategy extended into educational territory that aligned directly with the brand's allergen transparency positioning. Posts explaining the difference between gluten-free certification levels, or detailing how cross-contamination is managed in food production, attracted a highly engaged readership among people managing coeliac disease and food allergies. These readers were not casually health-conscious — they were actively searching for trustworthy food brands, and the educational content positioned this brand as expert and trustworthy before any commercial message was introduced.
Podcast appearances in months ten through twelve generated a measurable spike in subscriber growth — roughly 340 new sign-ups were attributed to two appearances on mid-sized UK food and wellness podcasts with audiences between 20,000 and 45,000 listeners. The brand offered podcast listeners a first-box discount code, which allowed attribution, but the more significant effect was sustained organic search growth in the weeks following each appearance. Podcast listeners who discovered the brand searched for it directly, and that branded search traffic converted at a substantially higher rate than cold paid social traffic. Using BlogFlow to consistently publish keyword-relevant content alongside podcast activity compounded the organic visibility effect over time.
Retention Mechanics Beyond Personalisation
Personalisation was the primary churn reduction driver, but the brand layered several additional retention mechanics that collectively kept monthly churn below 8% throughout the growth period. The most impactful was a box preview email sent four days before each delivery. Subscribers received a preview of their personalised selection with the option to swap up to two items before the box was packed. This single feature reduced cancellations in the week before billing by approximately 22%, because subscribers who disliked a particular item could resolve their dissatisfaction without cancelling entirely.
Pause functionality was introduced at month eight after churn data revealed that a significant proportion of cancellations occurred during holidays, house moves, and other temporary disruptions. Before the pause option existed, subscribers in those situations cancelled — often intending to resubscribe later but rarely doing so. After pause was introduced, 38% of subscribers who would previously have cancelled instead paused for between two and eight weeks. The reactivation rate from pause was 91%, meaning the pause feature effectively converted potential permanent churn into a temporary break with near-certain return.
Loyalty milestones were added at month fourteen, offering small rewards at three-month, six-month, and twelve-month subscription anniversaries. The rewards were modest — a bonus product, a branded tote bag, early access to new SKUs — but the psychological effect was meaningful. Subscribers approaching a milestone were measurably less likely to cancel in the two weeks before it, and post-milestone satisfaction scores were higher than the baseline average. Designing retention mechanics that work together, rather than relying on a single lever, is what sustains low churn at scale rather than just at launch.
Scaling Without Losing the Personal Feel
Growing from 120 to 5,100 subscribers introduced operational pressures that threatened the personalised experience. At low volume, the curation team could review edge cases manually — a subscriber whose allergen profile was unusual, or whose preference ratings were contradictory. At scale, those manual reviews became impossible without automation. The challenge was preserving the perception of individual attention while processing thousands of personalised boxes each week.
The solution combined automated filtering with a human review layer for flagged profiles — subscribers whose last three box ratings averaged below 3.5 out of 5, or whose skip rate exceeded 40%, were routed to a small customer success team for a brief personal outreach. This meant the team was spending human attention where it had the highest retention impact rather than distributing it evenly across all subscribers. The flagging system identified at-risk subscribers before they cancelled, rather than after, which changed the nature of the retention conversation entirely.
Brand voice consistency across all automated communications was equally important. Box preview emails, dispatch notifications, and allergen update alerts were all written in the same warm, knowledgeable tone as the brand's organic content — not in the generic transactional language of most e-commerce automation. Subscribers regularly cited email tone in qualitative feedback surveys as a reason they felt a genuine connection to the brand. At scale, maintaining that voice required a clear editorial guide and periodic audits of automated message templates, but the investment in tone consistency paid dividends in the subscriber relationships it sustained.
Try BlogFlow free at saltai.app — no credit card required.
SaltAI Team
SaltAI builds focused Shopify apps for food merchants and general merchants. Every app is tested in production at a real food store — including Vanda's Kitchen — before it ships.