He Left Deutsche Bank for a Camera and the Algorithm Could Not Tell the Difference

Hank had spent eleven years at Deutsche Bank making things that were not beautiful look presentable in slide decks, which is a skill that transfers poorly to anything outside of finance but which had made him very good at understanding what people needed to see in order to make a decision.

The photography had been the weekend thing. First his friends' weddings, then friends of friends, then strangers who had heard about him through the particular network that forms around people who are genuinely good at something and do not charge for it. By 2019 he had a portfolio that was better than most professionals he had seen and a day job that paid significantly better than photography and a growing awareness that he had these two facts in the wrong order.

He left in 2021. Not impulsively. he was not an impulsive person. He had modelled it on a spreadsheet over eight months, calculated the runway, identified the market positioning, researched the Shoreditch studio space. His wife said the spreadsheet was the most romantic thing he had ever done for her, which he took as a compliment and she meant as one.

The studio opened in March 2022. By September it was booking eighteen months out. The work was exceptional. not in a way that required anyone to say so, just in a way that was visible in the photographs themselves. Couples who found him by word of mouth booked him without hesitation. Couples who found him by other means sometimes asked about pricing first, which told him something useful about the different populations he was reaching.

The Google Ads campaign had been running since April 2022. He had set it up with the same analytical rigour he had applied to everything else. Keyword research. Competitor analysis. A/B testing on the ad copy. A budget that was proportional to his average booking value. A landing page that showed his best work in a gallery that loaded, on mobile, in six seconds.

Six seconds to load a photography portfolio on mobile is long enough for an engaged couple to decide the photographer is not serious. The gallery was the thing that would convince them. The gallery was loading after they had already left.

He had run the campaign for eighteen months. The return was positive but not proportional to the quality of the work or the investment in the setup. He knew, from the spreadsheet he maintained, that the word-of-mouth bookings converted at a rate four times higher than the paid bookings. He attributed this to trust. He was partly right.

The other part was six seconds on mobile in a market where his competitors' pages loaded in under two.

The Auditor's Take

Names and identifying details have been changed. The technical profile below is based on real observations from real sites. The diagnosis is always free.

Wedding photography is a high-consideration search with a specific mobile behaviour pattern. Couples search together, often on a single device, often in the evening, often with multiple tabs open comparing portfolios simultaneously. The decision to enquire is made in the first ten seconds of seeing the work. If the work has not loaded in ten seconds the decision is made for the wrong reasons.

A photography portfolio with high-resolution images served from a standard WordPress installation without image optimisation or CDN delivery will produce load times in the five to eight second range on mobile regardless of the quality of the images. The quality is irrelevant until the images load. They are not loading in time.

The Quality Score on the campaign reflected the load time. A landing page experience of Poor or Needs Improvement meant the campaign was paying a premium per click relative to competitors with faster pages. An analyst who had spent eleven years understanding premium pricing in financial instruments was paying a performance penalty on his own advertising because the infrastructure delivering his work did not match the quality of the work itself.

He called this character development. The Technical Tax does not care what you call it.

The fix is not rebuilding the portfolio. It is delivering the existing portfolio through infrastructure that serves it at the speed the market requires. The photographs do not need to change. The six seconds does.

Based on true events. All names changed. The Auditor's Take is a regular column on vsourcecode.com documenting real patterns from real sites. The diagnosis is always free.

Ready to Fix This on Your Site?

Run your site through PageSpeed Insights and send me the score. Free diagnosis. No pitch. Just the numbers.