They Don't Blame Their WiFi. They Blame You.
“They probably thought their internet was slow.”
It is the most common thing business owners say when slow-site feedback comes up. It sounds reasonable. It is the sort of charitable interpretation that makes sense from the inside. When your own page loads fine on your office connection, slowness elsewhere feels like someone else’s problem, maybe the user’s router, maybe their carrier, maybe peak-hour congestion.
The research says otherwise. Specifically, it says otherwise in ways that should change how seriously you take your load time.
The Attribution Question Has Already Been Studied
In 2010, Everard and McCoy published a peer-reviewed study in the Australasian Journal of Information Systems examining a direct question: when a website performs poorly, who do users hold responsible?
They ran scenario-based experiments where participants experienced site flaws, including degraded performance, and were either primed to attribute the problem to the website developer (internal attribution) or to the user’s own internet service provider (external attribution). Then they measured perceived site quality and trust.
Here is what they found, and it matters in two distinct ways.
First: trust in the website dropped equally regardless of which attribution participants made. Whether users blamed the site or blamed their connection, the trust damage was the same. The benefit-of-the-doubt scenario: “they’ll assume it’s their WiFi and come back”, simply did not hold in the data.
Second: when users attributed the problem internally (to the site and its operator), perceived site quality fell more severely than when they attributed it externally. Internal attribution added a quality penalty on top of the trust penalty.
Label this for what it is: HISTORICAL peer-reviewed experimental work, not an expiring benchmark. The year is relevant: 2010, but the mechanism it describes is a durable feature of how people assign cause. The finding does not expire when hardware improves.
Why the Attribution Has Only Gotten Stronger
The Everard and McCoy study was conducted when broadband penetration was still partial. In 2010, a meaningful portion of internet users shared an asymmetric DSL connection, experienced regular drop-outs, and had a lived reference point for “my internet is slow.” The rationalization “it must be their connection” had more cover then than it does now.
Today, your customer is on LTE or 5G, or on a home fiber connection. Their baseline experience is a world in which streaming video loads without buffering, maps render instantly, and restaurant bookings complete in under two seconds. When your page takes five seconds on their mobile device, the explanation “my internet must be slow” is not available to them in the same way. They loaded three other apps without delay in the last hour. The reference point is against them.
This is not a measured finding, it is a reasoned implication from the combination of the 2010 research and the documented improvement in consumer network speeds since. But it is a sound implication: the attribution to the website operator, already documented before broadband was universal, is more likely to occur today, not less.
What Users Actually Do in the Meantime
The attribution question is somewhat abstract. The behavioral consequence is not.
Forbes Advisor commissioned a survey conducted by OnePoll of 2,000 U.S. adults in 2024. One finding: when a page loads too slowly or is unresponsive, 48% of searchers simply visit the next website in search results.
They are not sitting there running a root-cause analysis on whether it was your server or their carrier. They are not filing a mental note to return when conditions improve. They are clicking back. The next result loads. The attribution question resolves itself irrelevantly because they are already gone.
This is the behavioral reality the attribution debate is arguing around. The average searcher is not choosing between “blame the website” and “blame the WiFi.” They are choosing between waiting and moving on. Most of the time, they move on. And when they move on, they go directly to a business that paid to appear in the same search results you did, and whose page answered first.
The Compounding Effect: When Slow Meets Imperfect
There is a second layer to this, from more recent peer-reviewed work.
Wells and Bullock published a study in the Proceedings of HICSS 2025, a peer-reviewed conference with an acceptance rate that reflects genuine vetting, based on an online experiment with 514 participants, using a 2×2 factorial design that isolated the effects of delays and presentation flaws independently and together.
The precise finding requires precision to state accurately: delays alone did not significantly reduce swift trust on average. The effect was conditional. What delays did was amplify the trust damage caused by presentation flaws, poor visual quality, interface imperfections, anything that a visitor might clock as amateurish.
Most SMB websites have some visual imperfection. An image that loads at slightly the wrong size. A font that renders inconsistently on mobile. A button that sits a pixel off where the spacing suggests it should be. None of these, individually, tanks a site. But when the page is also slow, the Wells and Bullock finding implies they compound: the visual imperfection lands harder because the delay made the visitor more critical, not less forgiving.
Slowness is not just a patience cost. It primes a visitor to judge everything else more harshly. A fast site is not just faster, it gets the charitable interpretation of the imperfections that every real site has.
A Tampa Bay Accounting Firm, Tax Season
Consider the following as unmistakably hypothetical.
A business owner is searching for a new CPA in late March. She opens three results from the same search page.
The first firm’s site takes about five seconds to render on her phone. She waits. The home page eventually loads. She taps the “Contact” button. The form takes a moment to respond. She is now running a mental calculation, if this firm’s website is this slow, what does their client communication look like? She does not complete the form.
The second firm’s site loads in under two seconds. The contact form responds immediately when she taps it. She fills it out and submits it.
The third firm’s site she never gets to.
At no point in this sequence did she consciously assign blame to anyone’s internet connection. She did not stop to wonder whether the first firm’s server was having a bad day. She formed an impression based on her experience, and the impression was: one firm runs a tight operation and one does not.
The first firm’s partners will review their monthly leads and note a slow March. They will not see this session in their analytics.
A Note on 5G as the Solution
Some owners hear this and conclude that as 5G adoption increases, the problem resolves itself. Better consumer connections mean faster loads, and faster loads mean the page gets there before the visitor leaves.
This misunderstands where the time goes on most slow SMB sites.
5G improves bandwidth, the rate at which data transfers. It does not reduce the round trips a browser must make to resolve DNS, negotiate security certificates, and download each resource in sequence. It does not reduce the time the browser’s main thread spends parsing and executing JavaScript. On most SMB sites built on WordPress, Wix, or similar platforms with accumulated plugins and third-party scripts, the bottleneck is not the pipe carrying the data. It is what the browser must do once the data arrives.
A faster consumer connection helps at the margins. It does not fix a site that ships 4 MB of JavaScript to a phone before rendering anything visible.
The Math Behind the Missing Leads
The 48% competitor-switch figure is a population average across 2,000 survey respondents. At a specific business scale it has a concrete shape.
Take a hypothetical Tampa Bay accounting firm with 50 organic search visits per month. If 48% of visitors who encounter a slow page leave without inquiring, per the Forbes Advisor/OnePoll 2024 survey of 2,000 U.S. adults, that is 24 lost contacts monthly. At a conservative $3,000 average new-client engagement and a 20% inquiry-to-client rate, those 24 represent roughly $14,400 in potential monthly revenue that never entered the pipeline. These are illustrative numbers at plausible SMB scale; the 48% behavioral rate is the sourced figure.
The Attribution Debate Is the Wrong Conversation
Whether your visitors blame you or their connection, the behavioral outcome is the same: nearly half of them go to the next result. Whether your delays compound your visual imperfections or not, the direction of the effect is the same: worse than if you were fast.
The rationalization “they’ll blame their WiFi” was never supported by the available evidence, and it is less supported now than it was when Everard and McCoy first tested it. The relevant question is not who gets the blame. It is whether your page loaded before the visitor decided it wasn’t worth waiting.
A free performance assessment from iServU answers that with real-user field data, what your actual visitors experienced on actual devices and connections, not a simulated lab test run from optimal conditions. Gabriel Lopez-Seco runs it directly. No account manager, no handoffs.
Schedule a free assessment at iServU.
Sources
- Everard, A. & McCoy, S. (2010). How Presentness of Information Technology Affects Its Perceived Usefulness. Australasian Journal of Information Systems, 16(2). https://doi.org/10.3127/ajis.v16i2.516
- Wells, J. & Bullock, G. (2025). Delays and Presentation Flaws: Effects on Swift Trust in Digital Contexts. Proceedings of HICSS 2025. https://doi.org/10.24251/HICSS.2025.531
- Forbes Advisor / OnePoll (2024). Consumer survey, n=2,000 U.S. adults, conducted by OnePoll on behalf of Forbes Advisor. Reported by Search Engine Land. https://searchengineland.com/people-wait-website-load-2024-stat-445223
