Photo by Pixabay
Originally Posted On: https://www.minneapolismade.com/web-design-speed-report/
We ran Lighthouse against 21 Minneapolis and Twin Cities web design agencies. The results are harsh. Only 4 of 21 passed Google’s Core Web Vitals threshold on mobile, 17 failed outright, and six agencies couldn’t paint their own homepage in under 20 seconds. The median mobile score was 55. For companies whose entire business is building fast, modern websites, that’s a credibility problem you can check in a browser in 30 seconds.
This post summarizes what we found and why it matters for any Minneapolis business about to hire a web design agency. The full leaderboard with all 21 agencies, every metric, and five branded charts lives on its own permanent research page.
Key Takeaways
- Only 4 of 21 Minneapolis web design agencies scored 80+ on Lighthouse mobile (our pass threshold)
- The median mobile score is 55, squarely in Google’s “slow” range per web.dev performance research
- 12 of 21 agencies exceed a 5-second Largest Contentful Paint, more than double Google’s “good” threshold of 2.5 seconds
- Minneapolis Made scored 98 mobile / 100 desktop with a 2.2-second LCP, the fastest LCP in the group
How does Minneapolis web design actually score on Lighthouse?
Only 4 of 21 tested agencies scored 80 or higher on Lighthouse mobile in our April 2026 benchmark. The median score was 55, which Google classifies as “moderate”. But mobile is the profile Google uses for ranking since the shift to mobile-first indexing in 2019, so moderate isn’t good enough for competitive local search.
Every agency was tested with Lighthouse v13.0.1 on April 16, 2026. Mobile profile used an emulated Moto G Power with slow 4G network throttling, the same conditions Google Search Console applies to field data. Desktop profile ran with no throttling. Median score reported per site. We’ve retained the raw dataset and republish the report quarterly; the next re-run is scheduled for July 2026.
Why does this matter for Minneapolis businesses?
Site speed isn’t a vanity metric. It controls how many visitors stay long enough to become customers, and how Google ranks you in local search. When your web design agency can’t deliver performance on its own homepage, that’s a preview of what your site will look like after they build it.
Google rolled Core Web Vitals into its ranking algorithm in 2021 and tightened the weighting in the March 2024 core update. The math is simple: a site that fails CWV ranks lower than a site that passes, all else equal. In a competitive local market like Minneapolis, “all else equal” is doing a lot of work. But even a small speed advantage compounds across every query you’re competing for.
Which Minneapolis agencies failed the benchmark?
Sixteen of the 21 agencies we tested scored below 80 on Lighthouse mobile. The worst performers had catastrophic LCP times:
The bottom three scores:
- Portkey SEO: 20 (LCP 32.1 seconds)
- Perrill: 31 (LCP 22.3 seconds, TBT 1,552 milliseconds)
- Iceberg Web Design: 32 (LCP 32.9 seconds)
These are agencies selling Minneapolis-area web design services whose own websites take more than 20 seconds to render main content on mobile. No amount of portfolio pieces or “years of experience” language changes what’s happening in a browser’s performance panel.
Unique Insight
The real story is the mobile-desktop gap. Eight of the 21 agencies score 90 or higher on desktop but collapse into the 50s or 60s on mobile. That gap exists because desktop has ten times the CPU horsepower, ten times the network bandwidth, and no thumb-sized tap targets. Google stopped caring about desktop scores for ranking purposes in 2019. Every agency still optimizing for the desktop test is optimizing for a machine nobody searches on anymore.
How did Minneapolis Made score #1?
We practice what we sell. The same techniques we deploy on client sites are visible in every byte of this one:
- Self-hosted fonts served as woff2 with latin-only subsets (no Google Fonts round-trip, no third-party DNS lookup)
- Inline SVG icons instead of a 30KB JavaScript runtime plus 73 API calls to fetch icon data on every page load
- LiteSpeed page cache at the web server level (near-zero Time To First Byte on cached pages)
- WebP images with automatic content negotiation via .htaccess (browsers that support WebP get it automatically, those that don’t fall back to PNG)
- Deferred analytics via requestIdleCallback (Google Tag Manager loads after the page is interactive, not before)
- Conditional CSS: blog pages ship 73% less inline CSS than service pages because they don’t need the process-timeline or shiny-CTA styles
None of these are exotic. They’re standard 2026 performance engineering that any competent web developer should deploy by default. The fact that most agencies don’t tells you something about their process.
Want to see your own site on this leaderboard?
We’ll run the same Lighthouse audit on your current site, compare your scores to every agency in the benchmark, and hand you a prioritized fix list. No sales pitch, no upsell.
What does a good Lighthouse score actually look like?
Google categorizes Lighthouse scores into three bands: 90-100 is “fast,” 50-89 is “moderate,” and 0-49 is “slow.” In our benchmark, only one agency (Minneapolis Made at 98) hit the fast band on mobile. Three more (Thrive at 87, The Mighty Mo at 85, Media Junction at 85) landed at the top of the moderate band. Everyone else was solidly mediocre or worse.
The underlying metrics matter more than the composite score:
- LCP (Largest Contentful Paint): When the main visual content finishes rendering. Good: under 2.5s. Poor: over 4s.
- TBT (Total Blocking Time): How long the page is frozen and can’t respond to taps. Good: under 200ms. Poor: over 600ms.
- CLS (Cumulative Layout Shift): How much the layout jumps around while loading. Good: under 0.1.
What does Google actually care about?
Mobile scores. That’s the short answer.
Desktop Lighthouse scores are a nice-to-have for users who browse on laptops, but they do not factor into Google’s ranking signals for any crawled URL.
That’s why the desktop-mobile gap is the most embarrassing data point in this report. An agency scoring 97 desktop and 61 mobile isn’t building for the web Google indexes. They’re building for a PDF preview.
Frequently Asked Questions
Why does site speed matter for local SEO?
Google has used Core Web Vitals as a ranking signal since 2021 and mobile page speed directly impacts bounce rate. A site that takes three seconds to load loses 32% of visitors before they see any content, according to Google’s own research. For Minneapolis businesses competing for local queries, speed is the difference between ranking and being invisible.
What counts as a good Lighthouse mobile score?
Google considers scores of 90-100 fast, 50-89 moderate, and 0-49 slow. Mobile is what matters because Google has used mobile-first indexing since 2019. Only 4 of 21 tested Minneapolis agencies scored 80 or higher on mobile in our 2026 benchmark.
How was this benchmark tested?
Every agency was tested with Lighthouse v13.0.1 on mobile (emulated Moto G Power, slow 4G throttling) and desktop. Tested on 2026-04-16. Median score reported per site. Raw data is available on request.
Will my agency be added to the report?
We re-run the benchmark quarterly with additional Minneapolis and St. Paul agencies. Email hello@minneapolismade.com to nominate a site for the next round. Inclusion is editorial and non-pay-to-play.
What’s the difference between LCP and TBT?
LCP (Largest Contentful Paint) measures when the main content visibly renders. TBT (Total Blocking Time) measures how long the page is frozen and unresponsive. Both matter: LCP controls whether a user sees your content; TBT controls whether they can interact with it.
How often is this report updated?
Quarterly. Next scheduled update: 2026-07-16. Historical results are archived with date stamps so readers can verify claims against the original dataset.
Why does Minneapolis Made’s site still show a 203ms TBT?
203 milliseconds is within Google’s “good” threshold of 200ms with a 3ms margin. The TBT comes almost entirely from our inline SVG icon rendering, a deliberate trade we chose over shipping a third-party icon library that would have added more bytes and external network requests. Compared to the 1,552ms TBT of the worst site in our benchmark, 203ms is effectively zero.
Related: The real cost of a slow website: revenue data you can’t ignore
Related: Core Web Vitals explained: LCP, TBT, CLS, and why they control your rankings
Related: How to choose a web design company in Minneapolis: what to look for beyond portfolio pieces
Related: Best web design companies in Minneapolis: ranked by verifiable data
Related: Minneapolis web design services: see how we earn the #1 speed ranking
Want the full benchmark with every chart?
The permanent research page has the complete 21-agency leaderboard, 5 branded charts, and a full technical breakdown of why the gaps exist.
