Background
Löplabbet is the specialty store Swedish runners go to when a pair of neutral trainers won’t cut it — the kind of shop where gait analysis comes before the checkout, and where the staff can tell a Saucony Endorphin from a Hoka Bondi without looking it up. Online at loplabbet.se, the catalog stretches from racing flats to recovery tools, and the audience stretches from marathon regulars to first-time 5K-ers.
Running retail has a particular quirk online: the long tail is very long. Every model has a dozen variants, every variant has a dozen use cases, and every use case comes with its own highly specific search queries. Demand doesn’t move evenly either — January resolutions, spring race season, autumn marathons, summer trail lulls. Covering that terrain by hand is a full-time job, and one that doesn’t scale.
Challenge
Before Amanda AI, Löplabbet’s ads were managed through an agency, and the results weren’t landing where the team wanted them to. Keyword mining, budget allocation, and the structure of the Shopping feed all need work. Constant manual adjustments were getting in the way of actually growing the account. Which new queries were worth bidding on? Where was spend going without a return? Which budgets deserved more? Answering those questions one at a time, through an agency, was slow.
The solution
Löplabbet handed its existing setup to Amanda AI on Google and kept the strategic direction in-house. Three areas moved first.
Keyword mining
The productive queries in running retail live out past the obvious ones: specific shoe models, foot-strike preferences, distance-specific training terms, and the language people use when they’re coming back from an injury. Amanda AI discovers new search terms continuously, and each one is run through an LLM that checks relevance against the target landing page before it’s added. A quality score threshold keeps the bar high, so the keyword set expands without turning into noise. Löplabbet allocates nothing to brand bidding, which means every cent of the mining work goes into growing the non-brand portfolio.
Budget allocation
Amanda AI redistributes spend daily across campaigns, and the advanced budget allocator reads more than just yesterday’s conversions. Underneath, the model separates seasonal effects, cyclical patterns, and longer-term trends, pulling from signals like day of the week, day of the month, paydates, and public holidays, against a 60-day performance window. Budgets are capped at a daily ceiling rather than chasing demand without a floor — the cheapest way to lose money in paid search is to let it run hot into a high-demand week no one checked. For Löplabbet, monthly levels shift with the business’s own rhythms, and extra budget goes in during specific campaign periods; the daily allocation takes care of itself.
Shopping feed structure
Amanda AI took ownership of Löplabbet’s full Shopping feed. Naturally, Shopping campaigns concentrate spend on products already proving themselves, which means a successful new running shoe might never get enough impressions to become one of the proven ones. Amanda AI deliberately spreads the budget wider across the catalog, forcing a broader set of products to get tested. Over time, that’s how a wider slice of the range actually ends up earning its keep.
“Using Amanda AI has genuinely been a lift for us. We’ve not only gained a higher conversion rate and an even better ROAS, but most of all, we’ve saved an enormous amount of time. When you see that the tool has made over 10 million optimizations, you understand how much time we’ve saved. Getting this kind of efficiency and uplift in results at a very reasonable cost makes us incredibly pleased.”
— Anton Lundin, Löplabbet
Result
Roughly a year in, the numbers tell an efficiency story rather than a fireworks one — which is usually the more useful kind:
- Revenue +29.47%
- ROAS +17.49%
- Conversions +10.84%
- Spend +10.20%
Revenue grew nearly three times as fast as ad spend. For a business with the seasonal swings of running retail — January resolutions, spring race season, autumn marathons — having budgets that automatically follow demand matters more than any single optimization. And for a team that used to sit between an agency and a spreadsheet, the most telling change might be the quietest one: the account is now mostly handled by the machine, and nobody has to go hunting for what’s working.
