Why Fixing Your Server Response Time Won't Fix Your Bad Reputation

I was recently approached with a digital brief for an ecommerce company. Decent-sized business, technical and marketing directors involved, clear budget in mind. The brief was straightforward: fix the website. Server-side rendering wasn't working properly, images were loading slowly, performance scores weren't great. It was the kind of technical brief that can deliverable work with positive outcomes. I knocked back the opportunity immediately.

The brief made sense on its own terms. But reading through it, one question kept coming up that nobody wanted to address.

What is AI saying about this brand right now?

That question changed the conversation entirely.

The Technical Objectives Were Legitimate, But Incomplete

Here is what the brief actually prioritised:

• Reducing website server response time
• Speeding up image loading and media delivery
• Improving overall performance scores and user experience

Each of these was a real objective worth pursuing. We all know that slow websites lose customers. Performance scores influence user satisfaction and faster load times have value.

The challenge was context. Looking at the full picture, those three objectives were addressing symptoms rather than the adderessing the underlying issue.

The website had text presented as images, not editable copy, not indexable content, but actual images of words and they didn't even use alt text. I told the potential client that this had been an issue in search since the late nineties and I was still struggling why designers still make uncrawlable websites. It was like going back to the stone age.

It was all the standard stuff that an experienced SEO consultant would look at. Google couldn't read the images. Neither can any large language model trying to understand what the business does, who it serves, or why anyone should consider it. The most basic requirement of a crawlable, discoverable web presence was simply absent. Getting a faster server response time was never going to resolve their real problem.

The website schema markup was either absent or some of it wasn't implemented correctly. On our 45 minute discovery session I explained to them that structured data would tell search engines and AI systems what a business is, what it offers, and how credible it is. The site was functional but devoid to AI system that now shape how people discover and evaluate businesses. I think they found it rather informative but possibly a little disturbing..

The Review Situation Needed Addressing First

Then there were the reviews. This was a sore point.

The site had reviews, most of which were fabricated. The genuine ones ranged from quite positive to genuinely concerning. No one was addressing the issue. The irony was, it was one of the few parts of the site had crawlable content. At least they could have chosen the bits that you may wanted to be hide from being crawlable. We got a laugh out of it but the penny was beginnning to drop.

As Google's systems are increasingly capable of detecting the authencities of reviews. This seemed to be one of the core issues. Reviews are cited from many sources, not exclusively from their own website.

I made them do that prompt ; "without searching the internet, what do know about xxx brand within the context of xxx", xxx being the core product they sold.

It was another one of those moments, everything went kind of quiet for a second and no one was laughing anymore.

What AI Was Actually Doing With The Brand Information

This is the shift that the brief didn't account for or even consider.

SEO used to mean ranking on page one of Google. That goal hadn't disappeared, but a significant AI layer sat above it now. When a potential customer types a question into an AI assistant, that assistant generates an answer. It may or may not include a link. It may simply state something directly. What it was saying wasn't great, it made me go and read the other reviews on other platforms. The problem is you can run but you can't hide from these issues.

I suggested it was an opportunity to demonstrate a bit of humility and tell the world how things were changing within the organisation for the better. Crawlable content on their website and conversations on other platforms would be the best soluton going forward.

As the website content was locked inside the images, the AI found little to cite. They were using a custom stack ( I have no idea why) and most of the schema was missing, the AI had no structured signals to validate what your business actually did.

Between that and the dodgy reviews, the technical performance of the server became secondary at that point... or did it?

Why the Scope of the Brief Mattered

Development of new website functionality was explicitly out of scope. Copywriting and creation of new content were out of scope. That meant the two areas that would have directly addressed the visibility and reputation issues were off the table before the work began.

Fixing server response time without adding indexable content seemed like a wast of time to me. And the underlying problems would stay exactly where they were.

That outcome didn't feel like a useful way to spend budget, which is why I took the conversation took a different direction.

What Happened Next?

The opportunity is in front of them. The technical and marketing directors involved are now reconsidering the situation for a later date, but I think that reassessment opened up something that was genuinely worth pursuing.

The budget was real. The commitment to improving the website was genuine. Redirecting some of that energy toward content structure, schema, and review management would have produced outcomes that the technical improvements alone simply couldn't deliver. But I didn't get the job and they went with a web dev team to deliver the original brief. It was worth a bash.

Sometimes the most useful thing a conversation can do is widen the frame before the work begins. They may come back to me at some point in the future, it wouldn't be the first time and it wont be the last. I wish them all the best, let's see how they get on.

Three Questions Worth Asking Before Any Brief Is Written

I asked Claude to use my LLM to create three questions summarising the copy of this blog post. I think it did quite well and it saved me five minutes.

One: What does AI say about your brand when asked directly?

Open three or four AI assistants and ask them to describe the company. Ask whether it is trustworthy. Ask how it compares to competitors. The answers reveal the reputation situation in real time, without any filtering.

Two: Can your website content actually be read and indexed?

If homepage copy exists as images, or if key service descriptions sit in formats that crawlers cannot parse, then the site is not competing in organic search or AI citation regardless of how quickly the page loads.

Three: Does your structured data tell the story you want told?

Schema markup for business type, services, reviews, authorship signals, and entity relationships is the infrastructure that allows search and AI systems to represent a brand accurately. Without it, those systems fill the gaps using whatever they find elsewhere.

Thanks Claude.