Do I need a new website, or should I fix my ads first?

On first calls, “rebuild” is the word that makes people flinch, and they’re right to. One prospect this year had just paid for a new website when I first spoke to him. He didn’t need another one. When I say rebuild, I mean something specific, so here’s how I decide what actually needs work.

Step one: check that the numbers are real

Before blaming the site or the ads, confirm that sales and leads are being recorded. A surprising share of accounts I open are optimizing on broken data: a form that stopped firing when a button was renamed, two analytics installs counting each sale twice, calls that never reach the ad platform at all.

Each of those is from an account I took over:

• A contractor’s site recorded zero website conversions after a button label changed from “Send” to “Submit.” The ads looked like they had stopped working. They hadn’t. The case study shows the fix.
• A furniture brand had two analytics installs on one property competing with each other, and its revenue showed up as unassigned. One install removed, one character changed, reporting back. That’s in the unassigned revenue case.
• An online store’s Google account had a second purchase action counting alongside the main one for weeks, so about one in eight claimed purchases was a duplicate. The numbers are in which platform drove the sale.

All three would have made a new website look necessary, and a tracking fix solved each one.

If tracking is broken, both the site and the ads look worse than they are, and any rebuild decision made on those numbers is a guess. Fixing it is a defined project, typically $2,500 on my pricing page, and it comes first because every later decision reads from it.

Step two: find where people drop off

With trustworthy numbers, the question gets simple.

The ads are the problem when: the search terms report is full of queries you’d never sell to, the cost per lead is rising while the site’s conversion rate holds steady, or the account is spending on placements and audiences nobody chose.

The site is the problem when: the right people arrive and leave. High-intent searches, paid clicks from people who should buy, and the page loses them: it’s slow on a phone, it doesn’t match what the ad promised, or the form or phone number is hard to find.

Most of the time it’s a little of both, which is why the fix is usually narrower than a rebuild.

When to keep the website you have

Keep it if it loads fast on a phone, the tracking works, and you can add pages to it. In that case the fix is usually a landing page or two built for your ads, not a new site. A paid visitor should land on a page that answers exactly what they searched, with one clear next step. That can sit on your current site.

Rebuild when the site can’t be fixed in place: it’s slow because of the platform underneath it, nobody can edit it, or it’s missing the pages that searches and AI answers need. Even then, carry every old URL across with a redirect so you don’t throw away what search engines already know about you. On a 224-page law-firm rebuild I checked every legacy URL live after launch, and 223 of 223 resolved (case study). On a medical practice’s move, I inventoried 533 old URLs before touching DNS and cut the site and its ad account over together (case study). That inventory is the part of a rebuild people skip, and it’s the part that protects the leads you already get.

Why rebuilding an ad account is usually the wrong call

Starting an ad account over feels clean, and it’s expensive. On Google, the account’s conversion history is what Smart Bidding learns from. Open a new account and you pay for weeks of clicks while it learns again, which is spending money to buy data you already had. Meta keeps most of its learning in the pixel and dataset rather than the ad account, so the bigger risk there is replacing or breaking the pixel during a site rebuild.

The better move is to restructure inside the existing account: cut the waste, fix the conversion goals, tighten the targeting, and keep the history. After a meaningful bid or structure change I leave it alone for about two weeks before judging it, because the platforms need that long to settle.

Before your next call

Ask whoever is proposing the rebuild to show you the number that proves it’s needed. If they can point to tracking, search terms or drop-off on a specific page, it’s a diagnosis. If they can’t, it’s a sales pitch.

My free audit checks tracking, the ad account and the landing pages together and tells you which one to fix first. Request it at /audit-request.