Website Redesign in New York

Because they are individually unimpressive and collectively the entire strategy, which makes them the obvious thing to consolidate in every rebuild. New York search is neighbourhood-level and frequently narrower, so a competitive site carries several location pages that each generate modest traffic while together covering the addressable market. Reported as a site-wide list they look like low performers. They are also the most expensive content in the country to rebuild — producing a genuinely specific Park Slope or Astoria page takes real local knowledge, and re-ranking it means competing again in the most contested market anywhere. The protection is a preservation audit segmented by neighbourhood with conversions attached rather than sessions, so each page's contribution is visible before anyone proposes merging them.

By recording the mobile conversion path template by template before the redesign and comparing it afterwards, which almost nobody does. New York decisions happen on a phone within minutes, so the specific things that carry conversions — a tappable phone number in the header, a short path to action, information reachable without scrolling past brand content — are also the things a cleaner design tends to move. The loss then shows up as traffic holding while calls fall, which gets attributed to the market. Practically that means setting conversion tracking before the rebuild so a baseline exists, documenting where the phone number and primary action sit on each template, and treating any change to those positions as a decision requiring justification rather than a styling preference.

By setting a performance budget at kickoff and testing under the conditions that actually apply, which are worse than any office test. A large share of New York traffic arrives on phones underground with intermittent signal, on congested street networks, or while walking — so a rebuild that performs acceptably on office wifi can be functionally unusable where the decision is made. The budget means agreed ceilings on page weight and on JavaScript shipped before interactivity, measured on every build. The testing means loading key templates on a throttled connection on an actual phone rather than accepting a Lighthouse score. New designs are almost always heavier than what they replace, so without explicit enforcement the rebuild reliably ships slower, and in this market slower means fewer calls.

Yes, and separating the two branches explicitly is worth doing during a redesign because it is otherwise the moment they get merged. Buffalo, Rochester, Syracuse and Albany behave as conventional mid-size metros where one substantive city page per market is the right unit, while downstate needs hyperlocal granularity. A rebuild that applies a single location template across both either fragments the Upstate markets into neighbourhood pages nobody searches for, or flattens the downstate hyperlocal tier into city-level pages that cannot rank in a walking-distance market. Both outcomes are common because a single location template is the default deliverable. Designing two branches with different templates and different parents costs one additional template and prevents a structural mistake that would otherwise persist for years.