360: Idiotic Mistakes We’ve All Made As Ecommerce Entrepreneurs With Kurt Elster

Podcast: Download (Duration: 54:38 — 62.8MB)

The most common ecommerce mistakes are fixable and mostly unglamorous: obsessing over page speed while your copy is one line long, driving paid traffic to a page that was never ready to convert, installing dozens of apps hoping the right combination fixes things, and guessing who your customer is instead of asking them.

In this episode I sat down with Kurt Elster of The Unofficial Shopify Podcast to compare the mistakes we each see most often across our stores and clients.

Below is the full list: why page speed anxiety is misplaced, what actually belongs above the fold, why nobody talks to their customers, app roulette, and chasing revenue instead of profit.

Key takeaways

• Core Web Vitals is a tiebreaker, not a ranking overhaul. Most large Shopify stores score in the low double digits.
• Copy beats design and speed. Fix positioning and story before optimizing anything technical.
• Hide zero-review widgets. A visible zero is negative social proof.
• Never send ad traffic to your homepage. Or to an unoptimized product page.
• Call your abandoned cart customers. Roughly 7 in 10 will talk to you.
• Deleting a Shopify app leaves its code behind in your theme.
• Split testing needs about 100,000 visits a month plus a 1% conversion rate.
• Chase profit rather than revenue. Sales at a bad ROAS feel like progress and are not.

Mistake 1: Obsessing over page speed

Core Web Vitals functions as a tiebreaker between two sites competing for the same placement, not as a wholesale ranking overhaul.

Kurt’s frustration is the anxiety it generates. Merchants are tearing their hair out over a letter grade while genuine problems on their sites go unaddressed.

The data supports his position. Looking across the Shopify stores he has partner access to, very few score above 30, and his seven, eight, and nine figure clients typically sit around 12. The stores scoring above 60 are ones built five to ten years ago that were never touched again.

Brian Dean’s research at Backlinko across a large sample found very little correlation between Core Web Vitals scores and user experience metrics, except for sites that were badly broken.

My own view is that the scores are achievable and expensive. All my sites are in the nineties now, and it took three weeks of working directly in the code, which is not a reasonable expectation for a non-technical owner.

The causes are consistent: apps, JavaScript, and failing to defer that JavaScript.

My theory on Google’s motivation is advertising. Cumulative Layout Shift specifically measures whether the page jumps while loading, which is exactly the behavior that produces accidental ad clicks, and a faster page produces more legitimate ones.

Mistake 2: Driving paid traffic to a site that is not ready

People conclude that Facebook or Google ads do not work when the actual problem is the page the traffic lands on.

The pattern is consistent. Someone reports that ads failed, and the site they were sending traffic to was never going to convert regardless of how good the ads were.

The two worst destinations are the homepage, which Kurt says should never receive ad traffic under any circumstances, and a standard unoptimized product page with a one-line description.

Effort goes into the ad copy while the destination gets none, which is backwards.

Mistake 3: Displaying zero reviews and zero shares

A review widget showing zero stars is negative social proof, and it tells visitors nobody has bought this.

The fix requires customization, since most review apps do not hide themselves automatically when there are no reviews. Kurt considers that an obvious feature these apps should ship with.

Social share buttons are a related problem. Kurt strips them from client sites by default, because heat mapping a best-selling product typically shows zero or one click.

The exception is Pinterest in the right vertical. Wedding is a perfect fit, which is why my store keeps the Pin It button and little else.

If someone pushes back, the answer is data. Heat map it or split test it, and the result is reliably the same.

Mistake 4: Weak positioning and no story

Copy is the foundation, and it consistently loses to design and technical work because writing feels like homework.

Kurt’s hierarchy starts with clarity above the fold. Within seconds a visitor needs to know what you sell and why, usually through a logo plus a three to five word tagline.

That clarity does the sorting for you. A visitor either wants to know more or recognizes this is not for them, and both outcomes are useful.

Story comes next, and it belongs everywhere: the homepage, the About page, and individual product descriptions.

Twirly Girl is the example we both point to. Michael Jamin is a genuine Hollywood screenwriter who wrote for King of the Hill and Beavis and Butt-Head, and he applied that craft to a video about his wife’s story that drives most of their sales.

Most people are not professional screenwriters, and the accessible version is a genuine About page. Something personal, so you read as a mom-and-pop operation rather than a faceless company.

The competitive logic matters here. Conscious consumerism means buyers want to know who receives their money, and a story is the one thing Amazon and Walmart cannot replicate.

Why the About page matters more than people think

The About page is typically the second or third most visited page on a store, contrary to the common assumption that nobody reads it.

Kurt’s framing makes the reason obvious. Someone stepping out of a parking lot asking for your credit card is insane, and a direct-to-consumer store is functionally doing that, so people need to know who is asking.

Session recordings show the behavior directly. A meaningful number of visitors add to cart, go read the About page, then either return to checkout or leave, which means that page is where the purchase decision resolves.

Mistake 5: Not talking to your customers

Almost nobody asks their customers anything, and most owners are working from a confident guess.

Kurt’s test is simple. Ask why customers buy, get an answer, then ask how they know. The answer is usually that they just know, which means it is a guess that might be right.

Three ways to get real answers:

Poll strangers. Services like PickFu return roughly 50 unbiased responses in about 15 minutes. Ask whether they would buy from this site assuming they liked the products. Friends will not tell you the truth, and strangers will.

On-site surveys. Hotjar incoming surveys catch immediate unfiltered reactions. Weirdos come with the territory, and eight people saying the same thing is a signal.

Phone calls. I have an abandoned cart script that emails me the customer’s phone number, and I call within minutes.

The script for the call: introduce yourself as the owner, say you run the business with your wife, mention you noticed they did not complete checkout, and ask for honest feedback. Lead with the compensation if it feels awkward, offering a large coupon or the product free.

Roughly seven out of ten people will talk to you.

What those calls actually reveal

The objections are specific and would never have surfaced otherwise.

When we launched a mother-daughter apron line, one caller explained she could not tell whether it would fit her daughter. We listed sizing by age range, and her child was outside the typical percentile for her age.

The measurements were on the page. What she wanted was a sizing table by height rather than age, because age is ambiguous.

Kurt’s addition is that sizing is the number one apparel objection, followed immediately by return policy, since buyers feel they are gambling. Our complication is that personalized aprons cannot be returned, which was another objection we only learned by asking.

Another failure mode surfaced the same way: someone abandoning because checking out on mobile was inconvenient, which points directly at needing an express payment option.

Mistake 6: App roulette

Installing app after app hoping the right combination will fix conversions makes the site slower, more cluttered, and more confusing.

Kurt’s description of the result is recognizable. The page takes a while to load, then spin to win, then enter your email, then another popup.

The self-defeating part shows up in his own data. Exit intent surveys on cluttered sites return complaints about the number of popups.

A related waste is apps that exist only to insert one line of JavaScript, because the owner does not want to touch theme files. That is roughly $8 a month forever for something a single edit would accomplish.

Why deleting a Shopify app does not remove its code

Deleting an app severs the developer’s access to your store and leaves any theme code they added exactly where it was.

That is a deliberate security behavior rather than a bug. You wanted them out, so they are out immediately.

The consequence is accumulation. Kurt quotes former Shopify CTO Jean-Michel Lemieux: websites get slow slowly. Old code from apps you forgot about two years ago still loads on every page view.

His own apps send a transactional email on uninstall asking why, explaining exactly how to remove the theme code, and offering to do it.

Finding it yourself is usually manageable. Most of this code sits at the top or bottom of theme.liquid, and the include line typically names the app, so you can comment it out or delete it.

The broader point Kurt makes is that basic HTML and a little Liquid put you in a dramatically better position. Your theme is your full-time salesperson, and you should be able to speak its language well enough to add tracking codes, verify a domain, and talk to any developer you hire.

Mistake 7: Guessing who your customer is

Writing all your copy for an imaginary customer without data is common, and the imaginary customer is usually yourself.

My own example: I assumed everyone buying from us was a wedding customer, until Facebook demographic data showed a large segment over 55.

Those were not people getting married. They were handkerchief collectors and crafters, mostly older women.

The fix was structural. Every landing page used to be about weddings. Now weddings have their own section, and the rest addresses embroidered blanks for crafting and collectors.

Neville Medhora’s example is the sharper version. He sold fingertip lights for raves and discovered plumbers were buying them to see under sinks, so he wrote copy aimed at plumbers and it converted well.

The lesson is that products do not have to stay in their intended use, and you only find that out by asking.

How to identify your most valuable customers

Look for anomalies in order size, then call those people.

Our numbers: repeat customers are 12% of buyers and 30% of sales. Average order value is around $60, half our customers spend under half of that, and roughly 10% spend double AOV while producing 50% of revenue.

With a small team, that math determines where attention goes. Many of our cheapest customers came from Facebook, so we redirected effort toward the large repeat buyers, which we call whales.

The identification process is manual and effective. We watch for unusually large orders, call them, ask whether they are a planner or buying in bulk, and offer a dedicated coupon code plus a named representative who handles their orders and ensures on-time delivery.

Kurt’s observation is that most people set up a wholesale program as a link on the site and wait. Being proactive is what converts these into customers for life.

The phone call is what does it. At that point you stop being a brand and become a person helping them.

What to ask in a post-purchase survey

Ask what they are using it for, and ask why.

The what is a product suggestion. The why reveals intent, which is what most people never scratch at and what actually changes your copy.

Our specific version asks whether the purchase is for business, personal use, or a gift. Anyone answering business gets a phone call, because they have identified themselves as a potential whale.

The rest feeds review requests, testimonials, and photos for social.

When split testing is worth it

Split testing needs roughly 100,000 visits a month and at least a 1% conversion rate before it produces conclusive results.

Every test needs a minimum of two weeks, which is frustratingly long while probability-to-be-best flips back and forth.

The reason most tests come back inconclusive is what people test. Design matters far less than copy, content, offer, and traffic quality, so design tests produce noise.

Two improvements Kurt recommends: use revenue rather than conversion as your primary goal, which folds in average order value, and limit how many tests and variants run simultaneously.

The tests genuinely worth running are offers. Finding your optimal free shipping threshold has real revenue impact, and tools like Ship Scout make it testable for Shopify Plus merchants.

Below that traffic threshold, use PickFu and heat maps instead. Heat maps, scroll maps, and movement maps are heuristic tools that work at any traffic level.

One student was determined to keep a sidebar that compressed their category page images. The heat map showed nobody clicking it and nobody even looking at it, and that settled the argument.

Mistake 8: Chasing revenue instead of profit

Sales arriving at a terrible return on ad spend feel like progress, which is what makes this mistake persistent.

Kurt’s nuance is that deliberate loss-leader acquisition is a legitimate strategy for brands spending investor money and planning to recover it on subsequent purchases.

The problem is doing it accidentally, or doing it while bootstrapped with thin margins and telling yourself you will iterate into profitability without deciding how long that gets.

The mistake underneath all of them

Kurt’s final answer is not starting, and mine is refusing to do the unglamorous work.

His prompt for getting unstuck: ask what someone smarter than you would do, then notice you already know the answer and go do it.

Mine is that people want to build an online business and do everything online. Picking up the phone, talking to someone face to face, and being personal are still the job, exactly as they are in a brick and mortar business.

Frequently asked questions

Do Core Web Vitals really affect rankings?

They act as a tiebreaker between sites competing for the same position rather than as a major ranking factor. Most seven to nine figure Shopify stores score in the low double digits, and research found little correlation between scores and user experience metrics.

Should you show review widgets with zero reviews?

No. A visible zero is negative social proof telling visitors nobody has bought the product. Most review apps do not hide themselves automatically, so this requires a small customization.

Are social share buttons worth keeping?

Usually not. Heat mapping a best-selling product typically shows zero or one click on share buttons. The exception is Pinterest in visual verticals like weddings, where the Pin It button genuinely performs.

How do you get honest feedback about your store?

Poll strangers rather than friends. Services like PickFu return around 50 unbiased responses in 15 minutes, and on-site surveys through Hotjar capture immediate reactions. Calling abandoned cart customers directly works too, with roughly seven in ten willing to talk.

Does deleting a Shopify app remove its code?

No. Deleting severs the developer’s access and leaves any code they added in your theme. That code keeps loading on every page view. Most of it sits in theme.liquid and names the app, so it can be found and removed manually.

How much traffic do you need to split test?

Roughly 100,000 visits a month with at least a 1% conversion rate, and every test needs a minimum of two weeks. Below that, use polling services and heat maps, which produce useful signal at any traffic level.

How do you find your most valuable customers?

Watch for anomalies in order size and call those buyers. Ask whether they are buying for a business, offer a dedicated discount code, and assign someone to handle their orders. The phone call is what converts them into customers for life.