Website Personalization for Agencies: Your Site, and a Service You Can Sell

An agency homepage must win fintech, healthcare, and ecommerce clients with the same hero — plus job applicants, partners, and competitors' interns. The standard fix (industry pages) only works if visitors find them. Personalization: the right headline per visitor, on the homepage, automatically.

Your own site: mixed traffic, matched headlines

Four buckets: prospects by industry, prospects unknown industry (fallback), job seekers, partners/press. Baseline agency: "We build brands and websites that grow businesses." Fintech prospect: "We've shipped 14 fintech brands through compliance review — websites that convert and pass legal." (subcopy: "Case study: how we took Northgate Bank from a 2011-era site to a 34% application lift, through three rounds of compliance review."). Healthcare: "HIPAA-conscious web design for clinics and health startups — patient trust, built in." Ecommerce: "Shopify and headless storefronts that lift conversion — our last 6 launches averaged +22%." Job seekers: "We hire slowly and keep people. See who's here and what we're working on." Fallback covers the rest; Prept never renders blank.

Why agencies should care beyond their own site

Every client site has the same mixed-traffic problem. The tooling was an enterprise contract or a dead Google product (Optimize, sunset September 30, 2023). Prept: one snippet + three data-prept attributes per client site — a ten-minute task you can charge for monthly.

The partner program

prept.dev/partners: 30% recurring commission per client, for as long as they stay. One snippet per client site — no shared accounts, no cross-client data. You keep the client relationship; Prept is infrastructure. Recurring, not a bounty: a client on the 50k pack ($129/month) = $38.70/month commission. Ten such clients ≈ $387/month for ten emails and ten snippets.

The margin math of the retainer add-on

Price the add-on at $600/month: headline/subcopy/CTA personalization on homepage + two top landing pages; variants written once (billable), system runs itself. Cost side for a typical SMB client (8,000 visits/month, 70% new, 30% returning, 3 elements): new 5,600×3 = 16,800; returning 2,400×3×0.25 = 1,800; total 18,600 credits/month. One 50k pack ($129) covers ~3 months → ~$48/month effective. You collect $600: ~92% margin before your time. Ten retainer clients: $6,000/month revenue on ~$480 credit costs. Honest version: month one is real work — price a $1,500–$3,000 setup fee; sell it as a monitored optimization layer, not set-and-forget. Check the 10% control holdout quarterly and refresh variants when positioning changes.

What to tell clients it isn't

"Is this A/B testing?" No — no fixed variants, no declared winner; 10% see the original as control. "Is this tracking my visitors?" Privacy-safe context only — referral, company-level firmographics, new vs. returning; no PII, no cross-site tracking. "What if it shows the wrong headline?" The fallback — original text as default; thin-profile visitors see the original; nothing ever renders blank.

Where to start

1. Put Prept on your own site first — you're your own case study. 2. Pick one existing client with mixed traffic; offer the add-on with a setup fee. 3. Join the partner program at prept.dev/partners from client one. 4. After two clients, make it a standard proposal line item. Agencies that win sell outcomes, not deliverables: "Your homepage now speaks to each visitor's industry, and here's the control-holdout delta."

Further reading: The Google Optimize Alternative — Google Optimize Migration Guide — A/B Testing vs. Personalization — Headline Personalization for SaaS. CTA: Install the snippet →.