Design approval is the last cheap moment in a website project.
Everything before it is a conversation. Everything after it is a change order. And a surprising number of the decisions that determine whether the finished site can ever be found in search are made, silently, in the design phase, by people whose job is not search.
These thirteen checks take about an hour. Each one costs nothing now and a rebuild later.
1. Does every page template have a place for a unique title?
Not the H1 on screen. The <title> tag, editable independently, per page.
If the CMS generates titles from a pattern, two hundred pages end up with near-identical titles competing against each other. Google resolves that contradiction by picking one and burying the rest, and it does not necessarily pick the one you wanted.
Ask: show me where I edit the title tag on a product page, and confirm it is separate from the heading.
2. Are the URLs decided, and are they stable?
URLs should be short, readable, and unchanged when a page gets renamed. /services/web-design survives a rebrand. /page-id-4471 tells nobody anything, and /2026/03/our-new-services-page-final will embarrass you.
Get the full URL map approved alongside the design. It is a five-minute document that prevents the single most destructive event in a website’s life.
3. Is there a redirect map for every existing URL?
If this is a rebuild rather than a new site, this is the most important item on the list.
Search engines rank URLs, not websites. Every ranking signal you have accumulated (external links, click history, topical relevance, age) attaches to a specific address. Change the address without a 301 redirect and the new page starts from zero while the old one returns 404.
Traffic losses of 40-70% are typical when redirects are missed, and the decline is gradual over three to six weeks, so by the time it is unmistakable nobody connects it to the launch.
Ask: where is the redirect map, and has it been built from both a full crawl and the Search Console URL list? A crawl alone misses orphan pages, which are frequently the ones holding external links.
4. Does content appear in the HTML, or only after JavaScript runs?
Google renders JavaScript, but on a delayed second pass, and complex client-side applications frequently render incompletely for crawlers. The failure mode is content that exists for users and does not exist for search engines: invisible to everyone working on the project, because it looks fine in a browser.
The check: load a page, disable JavaScript, reload. If the words are gone, the rendering strategy needs discussing before build, not after.
5. How heavy is the design going to be?
Ask for the estimated page weight of the homepage as designed. If the answer is “we haven’t thought about it,” that is the answer.
Google’s own case study data shows what weight costs. Agrofy Market’s 70% Largest Contentful Paint improvement produced a 76% reduction in load abandonment. Vodafone Italy improved LCP by 31% and recorded 8% more sales. Lazada improved LCP threefold and saw a 16.9% increase in mobile conversion rate. Nykaa’s 40% LCP improvement produced 28% more organic traffic from tier-2 and tier-3 cities.
Those numbers run in reverse when a redesign makes things slower, and a visually richer design with a slider, four font families and a dozen scripts routinely does exactly that.
6. Has anyone specified the image pipeline?
Images are the majority of transferred bytes on most business sites. The design phase is where this gets decided by default and badly.
Ask for: modern format (WebP or AVIF), compression, images served at display dimensions rather than scaled down in the browser, lazy loading below the fold, and the hero image preloaded. Also explicit width and height attributes, so the browser reserves space and the layout does not jump.
7. Is space reserved for anything that loads late?
Cumulative Layout Shift is usually explained as an annoyance. On mobile it is a conversion killer, because the user has already committed their thumb when the content moves and the tap lands somewhere they did not intend.
Redbus cut CLS from 1.65 to zero, halved Time to Interactive, and reported an 80-100% increase in mobile conversion rate. AliExpress improved CLS tenfold and saw a 15% reduction in bounce rate.
In the design: every ad slot, embed, banner and dynamically inserted element needs a fixed reserved space in the layout.
8. Can you put links inside body copy?
Contextual links from within paragraphs carry substantially more weight than templated navigation or footer links, and they are the most controllable ranking signal you have.
A CMS where editors can only link via navigation removes that lever entirely. Check that body content supports inline links with editable anchor text, and that the design does not make inline links visually indistinguishable from surrounding text, which is an accessibility problem as well as a usability one.
9. Is every important page within three clicks of the homepage?
Take the proposed navigation and count. Pages four or more clicks deep get crawled less frequently and start to fade from the index.
This is a navigation design decision, not a technical one, and it is far easier to fix on a wireframe than after two hundred pages have been built into a five-level hierarchy.
10. Where does structured data go?
Product, LocalBusiness, FAQ, Article, BreadcrumbList, whichever apply. These generate the rich results that make your listing visibly larger in the SERP.
If the current site has schema markup and the rebuild does not carry it forward, your listings quietly lose their star ratings, prices and FAQ expansions at launch. It is a common casualty of a redesign and nobody notices for months.
11. Has staging been checked for the two things that break launches?
Is staging indexable? If so, it will compete with production for its own content.
Will the staging noindex or robots.txt block ship to production? This is the one that takes an entire site out of the index within days of launch, and it happens more often than anyone admits.
Both are one-line checks and both belong on the launch checklist with a named owner.
12. Is the mobile design the real design?
Most design reviews happen on a large monitor with the desktop layout on screen, and the mobile version gets a cursory scroll at the end. For most businesses that is exactly backwards: the majority of visitors will only ever see the mobile layout.
Three things to check specifically on a phone-width mock, not a resized browser window:
What is above the fold? If the first screen is a full-bleed hero image and nothing else, the visitor has to scroll before learning what the business does. Every scroll loses people.
Is the primary action reachable with a thumb? Phone number, WhatsApp, booking button. In many markets a large share of commercial contact happens through messaging rather than forms, and a site whose only conversion path is a contact form is discarding enquiries from people who would have messaged.
Does the navigation work without a hover state? Dropdown menus that reveal on hover have no equivalent on touch. This gets discovered after launch with surprising regularity.
13. Who owns each item on this list?
The most common reason these checks fail is not disagreement. It is that nobody was assigned.
The designer assumes the developer handles redirects. The developer assumes marketing supplies the URL map. Marketing assumes it is technical. The result is a launch where nobody built the redirects and everybody thought someone else had.
Before sign-off, put a name against each of the thirteen items above. Not a team, a person. Items 3, 4 and 11 in particular are the ones that fall between roles, and they are also the three that can cost you the entire site’s search visibility.
This takes ten minutes in the design review and it is the highest-leverage ten minutes in the project.
The meta-point: bring the search conversation forward
Every item above costs minutes during design and weeks after launch.
The cheapest possible intervention in a web project is having whoever handles your search read the sitemap and rendering plan before the design is signed off. Not to write content, to say “that URL structure will lose you your rankings” while it is still free to change. Businesses that bring in an seo services nepal provider or their equivalent at architecture stage rather than three months post-launch consistently get better outcomes, and pay less for them.
Better still is a build team that treats this as their own remit. A studio doing web development in Nepal or anywhere that ships a redirect map, unique title fields, a rendering decision and an image pipeline as standard deliverables has removed the failure mode structurally, rather than leaving it to fall between the designer, the developer and the marketer, which is where it usually lands.
The one-page version
Print this. Walk through it in the design review.
- Unique editable title tag per page
- URL map approved and stable
- Redirect map built from crawl and Search Console
- Content present in HTML without JavaScript
- Estimated page weight, agreed
- Image pipeline specified
- Reserved space for late-loading elements
- Inline links supported in body content
- Everything important within three clicks
- Structured data carried forward
- Staging indexability and noindex checked
- Mobile layout reviewed at phone width, not a resized window
- A named owner against every item above
Thirteen items, about an hour. The alternative is discovering the problem in month four, when the answer is a rebuild.
