Written audit of routing, caching, Vercel cost, and migration path. Built for engineering leaders who need clear answers before committing to a 12-week rebuild.
- Audit
- Architect
- Scale
monthly retainer
Who this is for
You run engineering on a Next.js application — likely somewhere between Pages Router and App Router, or already on App Router but watching the Vercel bill climb. You want a senior outside opinion before you sign off on a rewrite or a big infrastructure change.
The pain today
- Team is split between Pages Router and App Router with no unified plan.
- Vercel cost is climbing faster than traffic, and nobody can explain why.
- ISR plus edge caching is misbehaving in ways the team can't reproduce locally.
- Leadership wants a written document before approving a 12-week migration budget.
- RSC payload is bloated and slowing Time to First Byte on key routes.
The outcome you get
- A written Next.js audit covering routing, caching, RSC payload, and Vercel cost.
- A phased Pages-to-App migration plan with hours, risks, and rollback paths.
- A Vercel cost model with concrete levers and estimated monthly savings.
- A clear rewrite-vs-refactor recommendation your leadership can act on.
- A zero-downtime migration sequence so users never see a degraded experience.
What the Next.js audit covers
The audit walks the codebase from top to bottom: routing model (Pages vs App vs hybrid), data-fetching patterns (getServerSideProps, RSC, client fetch, Server Actions), caching layers (Next.js Data Cache, Full Route Cache, Vercel CDN, any self-hosted Redis), middleware scope, ISR and PPR usage, RSC payload size on heavy routes, bundle size and client JS budget, image pipeline, metadata and SEO baseline, Vercel project settings (edge vs Node runtime, output modes, region selection), and observability.
Each finding lands in a table with severity, cost impact in Vercel dollars per month where measurable, and estimated fix hours. Nothing vague — every row has a number or a range.
A written Next.js audit covering routing, caching, RSC payload, and Vercel cost.
Imohub — what the advisory output actually looks like
Imohub needed to index 120k+ properties in Next.js with search latency under 0.5 seconds and a Google-friendly URL structure. The audit before the rebuild covered routing design (ISR plus on-demand revalidation), search backend selection (Meilisearch over MongoDB versus Postgres full-text), image pipeline (optimized formats, lazy loading, critical path), and deploy topology (Vercel plus AWS hybrid).
The written plan drove the rebuild. The result: under 0.5-second queries, top-3 Google rankings, and roughly 70% infrastructure cost reduction. The document came first. The code came after. That sequence matters more than most teams realize.
120k+: Properties indexed and searchable.
Zero-downtime Pages-to-App migration
The migration risk nobody talks about upfront: a big-bang cutover degrades the user experience before it improves it. I sequence migrations to avoid that. The App Router runs alongside Pages Router from day one. Static and marketing pages move first — low risk, fast feedback. Data-heavy list and detail routes move second, using RSC plus streaming. Authenticated flows and dashboards move last, where customer-component boundaries are most complex.
Each phase has hours, regression-test coverage requirements, a feature-flag checkpoint, and an explicit rollback path. Vercel's dual-routing support makes this practical even on large codebases. Teams on Next.js 15 or 16 also get a review of the async dynamic API changes (cookies, headers, draftMode) that quietly break routes during migration if not caught early.
Why Vercel bills spike and how I find the leak
The most common cost drivers I find in audits: middleware running on every request instead of a scoped matcher, edge functions invoked by static assets, ISR generating too many variants, and oversized RSC payloads that inflate bandwidth on routes with high traffic.
I build a cost model with three scenarios — current state, quick wins after small config changes, and the steady state after the full migration. Most teams find 20 to 35% savings in the quick-wins column before touching a single line of application code.
Pricing and scope
Advisory engagement at $5,499 per month, pro-rated for a 1 to 4 week scope. Deliverables: written audit report (15 to 30 pages), Vercel cost model with scenarios, prioritized fix list with hours, one 60-minute debrief call with leadership, and a follow-up call two weeks later.
If you want me to execute the fixes, the Applications subscription at $4,999 per month covers implementation. Most teams do the audit first and decide after. I take on 1 to 2 advisory customers per quarter by design, so the work stays focused.
Recent proof
A comparable engagement, delivered and documented.
Rebuilt a real estate portal at a fraction of the cost
Rebuilt Imóveis SC's real estate portal as ImoHub, a faster, more scalable successor, handling 120k+ properties with sub-second search and drastically reduced AWS costs.
Read the case studyKeep reading
Frequently asked questions
The questions prospects ask before they book.
When the decision has a cost or timeline tail that makes getting it wrong expensive. App Router migrations, Vercel cost spikes, and major architecture pivots are the three common triggers. Your team knows the product — I know the patterns that go wrong in those specific situations. An outside audit usually pays for itself in avoided rework, even if the team ends up executing everything.
1 to 4 weeks. A mid-size App Router codebase with reasonable test coverage takes 2 weeks. A larger Pages Router monolith with no automated tests and a complex ISR setup takes closer to 4. I give a firm estimate after a written scoping review of your repo, not before.
Read-only access to the Vercel dashboard and analytics is ideal. If that's blocked, I can work from a recent invoice and the usage graphs exported manually. The cost model is less precise without live data, but the categories of waste are usually the same regardless.
Yes. Next.js supports running both routers in the same project, which is the whole point of a phased migration. The audit produces a sequence that keeps Pages Router in production while App Router routes are tested behind feature flags. Users never see the transition until each phase is verified.
Yes. Edge vs Node runtime decisions, middleware cost analysis, and Server Action design are all inside the audit. Middleware that runs on every request is one of the fastest ways to inflate a Vercel bill. Server Actions that bypass caching on read-heavy routes are the other common issue I find.
That's the goal. The report is written for engineers and for leadership at the same time — each finding has a business-cost framing in Vercel dollars or engineer hours, not just a technical label. I've delivered these documents to CTOs at a $1B+ fintech unicorn and to early-stage startups; the format scales to both.