Holky s úspěchem - Coaching Platform for Women

Brief

Holky s úspěchem ("Successful Girls") is a coaching project by Soňa Stískalová and Simona Lukášová targeting women entrepreneurs and professionals aged 28–45. The product: the "Your New Self" workbook (one-shot purchase), follow-up coaching session and an Instagram community group. The site has to:

1. Explain the positioning (no "manifestation course", but concrete business outcomes)
2. Show real testimonials from clients (income increases, work-life balance, side businesses launched)
3. Sell the workbook through Stripe Checkout
4. Trigger a 5-day email onboarding sequence after purchase
5. Be mobile-first - the audience scrolls on iPhones in transit or in bed at night

Audience research

Before building anything I sat down with Soňa and Simona on five 30-minute calls with prior-cohort clients. Three key insights:

• Mobile dominance: 87 % of traffic to the coaches arrives via Instagram → mobile. Desktop is rare (only B2B purchases by corporate HR).
• Social proof > marketing copy: testimonials with concrete numbers ("I lifted my billing from 60K to 140K") convert 3× better than aspirational copy ("become a confident woman").
• Price anxiety: a 990 CZK workbook is right on the impulse-buy edge for the audience. Social proof and a money-back guarantee matter more than discount.

These three principles drove the entire design.

Mobile-first design

Not "responsive tweaks of a desktop layout". I started at 375 px (iPhone SE) and scaled up. Specifically:

• Hero photo full-width, headline overlaid, CTA above the fold
• Single-column linear scroll (no side-by-side desktop tricks)
• CTA buttons 44 × 44 px minimum (Apple HIG touch target)
• Sticky CTA bar at the bottom on mobile, none on desktop
• Testimonials carousel with continuous scroll-snap (no arrow buttons - mobile users don't tap them)

Resulting Lighthouse Mobile 96/100 Performance, 100/100 Accessibility.

Stripe Checkout for one-shot

This is the inverse case to Maruška. Here there's no subscription - it's a one-shot purchase of the workbook for 990 CZK. Stripe Checkout (hosted) is simpler than a custom form:

// src/app/api/checkout/route.ts
import Stripe from 'stripe';
 
const stripe = new Stripe(process.env.STRIPE_SECRET_KEY!);
 
export async function POST(req: Request) {
  const { email } = (await req.json()) as { email?: string };
 
  const session = await stripe.checkout.sessions.create({
    mode: 'payment',
    payment_method_types: ['card', 'bancontact'],
    line_items: [
      {
        price: process.env.STRIPE_PRICE_WORKBOOK!, // 990 CZK
        quantity: 1,
      },
    ],
    customer_email: email,
    success_url: `${process.env.NEXT_PUBLIC_URL}/uspech?session={CHECKOUT_SESSION_ID}`,
    cancel_url: `${process.env.NEXT_PUBLIC_URL}/zruseno`,
    metadata: { product: 'workbook-tve-nove-ja' },
    locale: 'cs',
  });
 
  return Response.json({ url: session.url });
}

After success it redirects to /uspech?session=..., where a Server Component fetches stripe.checkout.sessions.retrieve(id), shows confirmation, and triggers the Resend email sequence.

5-day onboarding sequence

After purchase, sending a workbook PDF and disappearing isn't enough. A user who paid 990 CZK and never returns to the workbook is lost lifetime value. I added a 5-day email sequence:

Implemented via Resend audiences + scheduled emails:

// from Stripe webhook checkout.session.completed
import { Resend } from 'resend';
 
const resend = new Resend(process.env.RESEND_API_KEY!);
 
export async function startOnboardingSequence(email: string, customerName: string) {
  await resend.contacts.create({
    audienceId: process.env.RESEND_AUDIENCE_ID!,
    email,
    firstName: customerName,
    unsubscribed: false,
  });
 
  // Resend Schedule API - delayed sends
  for (const step of onboardingEmails) {
    await resend.emails.send({
      from: 'Soňa & Simona <ahoj@holkysuspechem.cz>',
      to: email,
      subject: step.subject,
      react: step.template({ name: customerName }),
      scheduledAt: new Date(Date.now() + step.dayOffset * 24 * 60 * 60 * 1000),
    });
  }
}

Campaign open rate: 64 % (vs the 22 % consumer benchmark). List quality (just-paid users) + copy relevance + first-person sender name.

Testimonial CMS - zero overhead

Testimonials change occasionally - a new client, a new monthly Instagram screenshot. Instead of a headless CMS, I went with MDX entries:

content/testimonials/
├── jana-fakturace.mdx
├── eva-vlastni-firma.mdx
└── martina-burnout.mdx

Frontmatter:

---
name: 'Jana K.'
role: 'Marketing consultant'
result: 'Billing 60K → 140K in 4 months'
photo: '/testimonials/jana.jpg'
publishedAt: '2026-01-12'
featured: true
---
 
> "Before the workshop I couldn't ask for my price. Now I bill more for an hour than I used to bill in a day."

Clients send Soňa an Instagram screenshot, Soňa forwards to me, I make an MDX commit. No admin dashboard, no Sanity/Strapi licenses.

ISR for testimonial freshness

A static build is fast, but when Soňa adds a new testimonial we shouldn't have to rebuild the entire site for it. ISR handles it:

// src/app/[locale]/page.tsx
export const revalidate = 3600; // 1 hour
 
export default async function HomePage() {
  const testimonials = await getFeaturedTestimonials();
  return <Page testimonials={testimonials} />;
}

When I push an MDX commit, the Vercel build finishes in ~25 s, but the user sees the new testimonial after the next request past the hour (ISR revalidation). Trade-off: up to 1 h delay from commit to visible. For testimonials, fine.

A/B testing via Vercel

The hero photo has two variants - Soňa+Simona together vs. Soňa solo with a laptop. Vercel Edge Config + middleware:

// middleware.ts
import { NextResponse } from 'next/server';
import { get } from '@vercel/edge-config';
 
export async function middleware(req: Request) {
  const variant = (await get<'A' | 'B'>('hero_variant')) ?? 'A';
  const res = NextResponse.next();
  res.cookies.set('hero_variant', variant);
  return res;
}

After 14 days (A: 1,200 visits, B: 1,240 visits) measuring CTR on "Buy workbook":

• Variant A (both together): CTR 3.8 %
• Variant B (Soňa solo): CTR 5.2 %

Solo converted better - hypothesis: the "I see myself in her" effect is stronger when one person is in the hero, not two. Permanent rollout B.

Lessons

• Audience research first, design after. The hero photo would have ended up at variant B even without the test, had I just asked clients "where did you see yourself in an ad".
• Mobile-first is not mobile-only. The desktop version is half the UX work, but Lighthouse 96 mobile and 99 desktop both demand the same precision measurement.
• One-shot Stripe Checkout is simpler than subscription. No failed payments, no dunning, no retention dashboard. The workbook sells or it doesn't.
• The 5-day email sequence yields 64 % retention. Without it, the user pays, downloads the PDF, forgets. Upgrade lifetime value (another coaching session, eventually a cohort) is 4× the workbook revenue.
• Vercel Edge Config A/B testing is underused. 50 lines of middleware, no external service. Plenty for 1k–10k traffic.