Daily conversion rate swings are weather. Weekly trends are climate. React to the trend and ignore the Tuesday, because the Tuesday is mostly noise wearing a trend costume, and chasing it is how stores end up “fixing” things that were never broken.

Every merchant knows the ritual. Conversion was 2.4% Monday, 1.6% Tuesday. Panic. Was it the theme update? The new app? Did checkout break? An hour of tab-shuffling later, no cause is found, and Wednesday politely comes in at 2.5% as if nothing happened. Because nothing did. On a store doing a few hundred sessions a day, a handful of orders landing on one side of midnight or the other moves daily CVR by half a point. One wholesale-minded browser on a big traffic day drags it down. A single shared link in someone’s group chat props it up. None of this is signal.

How to read the board without fooling yourself

• Compare week to week, and like to like. This week against last week, this Tuesday against last Tuesday. Weekend shoppers and weekday shoppers are different populations, so a Saturday will always “beat” a Wednesday.
• Pick a review day. One fixed morning a week for reading CVR, AOV, and revenue per visitor as weekly numbers. Checking daily is allowed; reacting daily is not.
• Know your exceptions. A sudden move that coincides with a deploy, a price change, or a traffic spike from one source deserves same-day investigation. That is diagnostics, not trend-reading: you are checking whether something broke, not whether the store “got worse.”
• Annotate the calendar. Sale started, influencer posted, theme updated, email sent. Six weeks from now, during BFCM chaos, those notes are the difference between explaining a spike and mythologizing it.

The discipline matters double from here to Cyber Monday. Q4 traffic gets spikier, promo calendars get noisier, and the daily number gets even less trustworthy just as your anxiety about it peaks. The stores that hold their nerve read climate. The rest rebuild their homepage because of one rainy Tuesday.

Session replays show you what your analytics never will: the shopper who wanted to buy and could not figure out how. Your dashboards report that 400 people left the product page; a replay shows one of them rage-tapping a size chart that would not open, scrolling past a buy button that blended into the theme, and giving up at a shipping calculator that silently errored. Numbers tell you where the funnel leaks. Replays tell you why, in a way that is impossible to unsee.

Microsoft Clarity is free, unlimited, and installs on Shopify in minutes; Hotjar plays the same role if you prefer it. There is no budget excuse here. The only real cost is the discipline to actually watch, so here is the assignment, sized deliberately small: ten replays, this week, twenty minutes total.

How to watch ten sessions like a professional

• Don’t watch randomly, watch the wounded. Filter for sessions with rage clicks, dead clicks, or quick-backs, on mobile, that reached a product page and did not purchase. Those are your almost-customers.
• Watch five from your biggest leak. You found your worst funnel step back in tip 11. Pull replays from exactly that step and watch the leak happen in person.
• Log one line per session: what they were trying to do, where it went wrong, what you would fix. Ten lines is a roadmap.
• Count repeats, not incidents. One shopper misses the variant picker, that is an anecdote. Four of ten miss it, that is your next fix, and probably your next A/B test.

Two warnings from experience. First, replays are anecdotes, and anecdotes are gasoline: one vivid video of a struggling shopper can hijack your roadmap harder than a spreadsheet ever could. Use replays to generate hypotheses and your metrics to rank them, never the reverse. Second, expect humility. Watching real thumbs navigate the store you polished for months is the single most ego-bruising exercise in ecommerce. Shoppers skim past the banner you argued about for a week, tap on things that are not buttons, and use your carefully designed mega-menu never. That bruise is the value: every replay replaces an assumption with an observation.

Do this weekly through the season, ten sessions at a time. With BFCM traffic arriving in six weeks, every confusion you catch now gets fixed once and pays out on every one of the thousands of sessions that would have hit it in November.

Why Is There Bot Traffic in Your Shopify Analytics?

Because the bots that matter don’t announce themselves. Most analytics tools filter the ones that do: search engine crawlers say who they are, and they are easy to drop. The expensive bots are scrapers watching your prices, rendering services that load full pages in a real browser engine, monitoring tools and link checkers. They run JavaScript, fire page views, and look like a visit.

They also never buy. So every bot session lands at the bottom of every rate you care about. Sessions go up, conversion rate goes down, add-to-cart rate goes down, and a collection that bots happen to crawl hard looks like it is failing.

This came to a head in early October 2026. A merchant sent us a GA4 report showing a surge in sessions that went nowhere. We traced it to a single rendering farm: the same browser build every time, a screen of exactly 1800×1125, and an outer browser window of 1808×1210. That window is bigger than the screen it runs on, which a real browser can’t do.

It wasn’t one store’s problem. Because we see every store we serve, we could follow that one fingerprint across the network, day by day.

One merchant looking at their own analytics sees a traffic spike. Looking across the network, it is plainly one bot.

What Does Bot Traffic Do to Your Conversion Rate?

It sinks it. A bot session adds to the denominator and never to the numerator. Take one fashion store from that week: bots were about 30% of its sessions. With them counted, its collection conversion rate was 0.27%. With them removed, 0.39%. Its cart rate moved from 1.30% to 1.85%.

The same goes for every comparison you make. If one collection gets crawled twice as hard as another, its conversion rate looks worse for reasons that have nothing to do with the products on it. Removing bots makes collections comparable again.

What Changed in Dynasort Insights?

Every storefront session Dynasort measures now gets one of three labels, and Insights count the first two by default.

A fingerprint we identify as a bot today is also removed from the days it visited before we caught it, so Insights get cleaner after the fact, not only going forward.

How Does Dynasort Tell Shopping Agents From Bots?

By what they do at the cart. Shopping agents are new, and they are customers. When someone asks an assistant to find and buy a pair of running shoes, the assistant visits your store, browses a collection, and adds to cart on that person’s behalf. Filtering those visits out would hide real demand. So the hard part of this project wasn’t spotting automation. It was telling automation that shops from automation that doesn’t.

Our rule is short: agents shop, bots don’t. An agent exists to buy something for a person, so over time its traffic produces carts and checkouts. A bot never does.

A single store can’t apply that rule, because it only sees its own traffic. We can. Every day we group sessions by device fingerprint across every store we serve and compare what each fingerprint did with what it should have done at those stores’ normal add-to-cart rates.

The network test sits alongside four other signals:

• Declared identity. Crawlers that name themselves are bots. Assistants that fetch a page live for a person, and say so, are agents.
• Signed agent requests. Some agent operators sign their requests using Web Bot Auth, so we can check who sent them. A signature from a known shopping-agent operator marks an agent.
• Impossible browser geometry. A browser window bigger than its screen, a window with no screen, or the square screens headless browsers ship with. We allow for rotated tablets and the quirks of phone browsers.
• Hosting networks. People rarely shop from a data center. A session from a hosting network with almost no activity counts against it; the same session with a cart counts as a shopping agent.

One honest limit: an assistant that hides what it is and browses without ever buying looks exactly like a scraper, so we count it as a bot. It produced no shopper outcome, so leaving it out doesn’t hide a sale.

What Will You See in Insights?

Cleaner numbers, a line telling you how much was left out, and a setting if you want a different view. There is nothing to turn on.

• Your numbers count shoppers. Sessions, conversion rate, add-to-cart rate, exits and bounces now count people and shopping agents.
• A marker on your timeline. The day bot filtering started is marked, so a step up in conversion rate there is easy to read.
• A setting if you want a different view. Settings > Traffic in Insights offers Shoppers (the default), People only, or All traffic, with your last seven days by type shown before you pick.
• Sorting doesn’t change with the setting. Your sort order never used bot visits and never will. The setting only changes what Insights display.
• A flag for traffic that doesn’t add up. If a collection’s sessions at least triple week over week while its cart rate falls by two thirds or more, Insights show “Traffic jumped without carts following”. That is the shape of a bot we haven’t identified yet, and it is worth a look.

Does Dynasort Store Shopper IP Addresses?

No. To recognize hosting networks we look at where each visit comes from, and we built that so we never keep the address itself. From each visit we keep the country, the network owner (for example, a home internet provider or a hosting company), a truncated address, and a one-way hash of the full address made with a key that changes every day and is deleted soon after. That is enough to spot one machine sending thousands of sessions today, and not enough to follow a person from one day to the next.

Country and network data come from IP Geolocation by DB-IP, licensed under CC BY 4.0.

Frequently Asked Questions

Does Dynasort count bot traffic in Insights?

No. Since October 13, 2026, Dynasort Insights leave bot sessions out by default. Sessions, conversion rate, add-to-cart rate and every other Insights number count people and shopping agents only. You can switch to People only or All traffic in Settings > Traffic in Insights.

Does bot traffic affect how Dynasort sorts my collections?

No. Dynasort sorting never uses bot visits, whatever you pick in the Insights traffic setting. Product views, cart adds and conversion rate used for sorting always come from people and shopping agents.

What is the difference between a shopping agent and a bot?

A shopping agent is an assistant that browses and buys on behalf of a real person, such as ChatGPT’s agent. A bot is a crawler, scraper or page-rendering service that never buys. Dynasort counts shopping agents as shoppers and leaves bots out. The short rule: agents shop, bots don’t.

How does Dynasort tell bots apart from shoppers?

Dynasort sees every storefront it serves, so it judges each device fingerprint by its behavior across the whole network. A fingerprint with many sessions and zero carts, where those stores’ normal rates would predict many carts, is a bot. Dynasort also uses declared identities, signed agent requests (Web Bot Auth), browser windows bigger than the screen they run on, and hosting-network signals.

Why did my conversion rate go up in Dynasort Insights?

Because bots no longer pad the session count. Your orders and carts did not change; the number they are divided by got smaller and more honest. At one fashion store where bots were about 30% of sessions, collection conversion rate moved from 0.27% to 0.39%. The Insights timeline marks the day bot filtering started.

Are older Insights days filtered too?

Days from October 6, 2026 on are filtered. Earlier days stay as they were, and the Insights timeline marks when filtering began so the change is easy to read.

Does Dynasort store shopper IP addresses?

No. Dynasort never stores a shopper’s raw IP address. It keeps the country, the network owner, a truncated address and a one-way hash that changes every day. Location data comes from DB-IP Lite, licensed CC BY 4.0.

Can a real shopper be counted as a bot by mistake?

A person who adds to cart or checks out is counted as a shopper unless the visit also carries strong automation evidence, like a declared bot identity. The one honest gap: an undeclared assistant that browses and never buys looks exactly like a scraper, so it is counted as a bot. It produced no shopper outcome, so leaving it out does not hide a sale.

Zero-result search rate is the percentage of site searches on your Shopify store that return nothing, and we would bet you have never looked at yours. Merchants track conversion rate to two decimals, argue about hero images for weeks, and meanwhile a measurable slice of their highest-intent visitors are typing a question into the search box and getting a blank page for an answer. Not a bounce from a cold ad click. A shopper who told you, in their own words, exactly what they wanted to buy, and got nothing.

Back in tip 5 we established that searchers convert at a multiple of browsers. That is what makes this leak expensive: every zero-result search is drawn from your best-converting population. The same failure on a random page costs you an average visitor. Here it costs you a great one.

Find your number this week

Shopify tracks this for you: the behavior reports in your admin include top online store searches and, crucially, top searches with no results. Divide no-result searches by total searches for the month: that is your zero-result rate. Then read the actual failed queries, because the rate tells you the size of the leak and the queries tell you the cause. They sort into three buckets:

• Vocabulary misses. They typed “hoodie”, your catalog says “pullover fleece”. You covered the fix in tip 30: synonyms.
• Typos. “Sneekers” on a phone keyboard. Fat thumbs get a full tip of their own later in the series.
• Genuine gaps. Products you do not carry. This bucket is free market research: repeated demand for something you could stock is a merchandising roadmap written by your customers.

Whatever the bucket, the page itself should never be a dead end; a zero-result page with fallback products and suggestions recovers shoppers a blank page loses (that rescue playbook was tip 43). We built this measurement into Dynasort because of how consistently it surprises merchants: our search tracking counts every zero-result query per store, so the leak has a number and a name the first time you open the report.

Your search box is a suggestion box where people deposit purchase intent. Read it, count the blanks, and fix the three buckets in order of volume. Few metrics on your store connect this directly to money left on the table.

Search conversion rate vs browse conversion rate is a comparison most Shopify merchants have never run on their own store, and it is the single number that decides how much attention your search box deserves. Back in tip 5 we told you searchers convert at a multiple of browsers. Today you stop taking our word for it and measure your own multiple, because “searchers convert better somewhere on the internet” funds nothing, while “searchers convert three times better on my store” is a budget line.

How to get both numbers

In GA4, the built-in view_search_results event marks sessions that used site search. Build two segments, sessions with that event and sessions without, and read purchase conversion for each over the last 30 days (90 if your traffic is thin; you know your sample-size manners from tip 38). No GA4? A serviceable proxy: compare conversion of sessions that landed on or visited your /search page against the store average. Cruder, but the gap will still show.

Reading your result

• Big gap, low search usage: your highest-intent behavior is being rationed by a hidden search icon. Making the box prominent (tip 5’s VIP-lane treatment) is the obvious first move.
• Big gap, healthy usage: congratulations, you have a compounding asset. Synonyms (tip 30), search log mining (tip 17), and the zero-results work coming later in this series all pay out at the multiplied rate you just measured.
• Small gap: rare and interesting. Either your browsing experience is genuinely excellent, or your search is bad enough that it burns high-intent visitors down to browser-level conversion. Run five real customer queries from your search logs and you will know which within minutes.

One measurement, three diagnoses, and a permanent answer to “is search work worth it on our store.” Write the two numbers next to your tip 1 baseline; we will be using them again.

Add-to-cart rate is the leading indicator of your Shopify conversion rate, and if you only watch purchases you are reading yesterday’s newspaper. A purchase is the end of a long chain: the shopper had to be persuaded, then survive shipping costs, checkout friction, a payment form, and their own second thoughts. An add-to-cart is the moment of persuasion itself, captured at the source.

That makes it valuable for two reasons. First, it is faster: for every order you get, you collect several add-to-carts, so the signal firms up in days instead of weeks. Second, it is cleaner: it isolates the question “did this page sell the product?” from the separate question “did checkout let them buy it?” Those are different diseases with different cures, and blended conversion rate mashes them into one useless symptom.

How to use it

• Baseline it store-wide. Shopify Analytics reports “Added to cart” sessions on the conversion funnel. Note the rate next to the three numbers from tip 1.
• Then read it per page. The store-wide number is weather; the per-product number is a diagnosis. A high-traffic product with a weak add-to-cart rate is your best CRO target on the whole store, and you found it without running a single test.
• Judge page changes on ATC first. New photos (tips 3 and 32), rewritten copy, size guidance: watch add-to-cart rate for the verdict, then confirm on orders once enough time has passed.

One caution: add-to-cart is a leading indicator, not a currency. A “free gift, just add it” widget will juice ATC while doing nothing downstream. Use it to detect persuasion, and let revenue per visitor, the honest metric from tip 1, remain the final judge.

Conversion rate can lie to you, and in a series about conversion optimization that sentence deserves its own day. CVR is a ratio of orders to sessions. It says nothing about what those orders were worth, and any metric that ignores money can be improved in ways that lose money.

The classic examples are not hypothetical, they are the default failure mode of half the “CRO wins” you will read about:

• The blanket discount. 15% off sitewide lifts conversion, shrinks every order, and the celebration email goes out before anyone checks contribution margin.
• The free-shipping limbo. Drop the threshold low enough and conversion climbs while the carts that used to add one more item to qualify stop bothering.
• The impulse-price hero. Push your $12 accessory to every top slot and watch conversion rise as your average order sinks. More orders, same revenue, more shipping labels.
• The urgency popup. Panic-buying converts. It also selects for the smallest, most regretted, most returned orders on your books.

None of these are scams. Some are even the right move for the right store. The point is that conversion rate alone cannot tell you which, because every one of them books its win in CVR and hides its cost in AOV.

Revenue per visitor catches the trick

You baselined revenue per visitor back in tip 1: total sales divided by total sessions, conversion rate and average order value multiplied into one honest number. RPV is the scoreboard precisely because it cannot be gamed from one side. Juice conversion by shrinking orders and RPV shrugs; grow orders by scaring off buyers and RPV shrugs again. It only moves when a session genuinely becomes worth more.

So institute a two-line rule for every experiment, every app, every offer from here to Cyber Monday: report CVR, and report RPV next to it, split by device like tip 19 taught you. A change that lifts CVR and holds or lifts RPV is a real win. A change that lifts CVR while RPV falls is a discount with better PR, and you should treat it like one: sometimes worth running, never worth mistaking for progress.

Conversion rate is the applause. Revenue per visitor is the box office. Enjoy the applause, bank the box office.

Mobile vs desktop conversion rate is the first split you should make on any Shopify analytics number, because the blended figure you have been quoting is an average of two different stores. One store gets browsed on a couch with a thumb. The other gets bought from at a desk with a credit card already out. They share a URL and almost nothing else.

A blended conversion rate is like a blended body temperature for two patients. “Averaging 37 degrees” sounds healthy right up until you notice one of them is on fire. On most stores mobile brings the majority of sessions and converts at a fraction of desktop, which means the blended number is quietly dominated by your weakest experience while feeling reassuringly stable.

Make the split, then read it

In Shopify Analytics, open your sessions and conversion reports and break them down by device type; in GA4, add “device category” as a dimension to any funnel or conversion report. Ten minutes, and you now have two conversion rates, two AOVs, and a traffic share for each device. Write them next to the baseline you took in tip 1.

Some gap is structural and not your fault. People genuinely research on the phone and buy on the laptop later, and no sticky button will change that. But the size of the gap is yours. As a working rule: if mobile converts at less than half of desktop, you have a mobile problem, not a “customers prefer desktop” anthropology finding. Something in the phone experience, speed, layout, forms, reachability, is charging you a toll on the majority of your visitors.

The split also changes how you read every experiment for the rest of this series. A change that lifts mobile 10% and does nothing on desktop can drown in a blended number. From now on, every before-and-after you run gets judged per device. Blended numbers are for investor decks. Split numbers are for fixing things.

Funnel step-off analysis answers the only question that should decide your next CRO project: where, exactly, do my visitors give up? Every store’s funnel has the same five steps, session to product view to add-to-cart to checkout to purchase, and every store leaks at every step. But one step always leaks worst, and fixing anything else first is redecorating a house with a burst pipe.

This is the discipline most CRO advice skips, including, frankly, series like this one. The next 89 tips are a toolbox, not a to-do list. Whether the badge tip or the checkout tip matters for your store depends entirely on where your funnel bleeds, and that is knowable, today, from data you already have.

How to run it in 20 minutes

Shopify’s analytics reports give you sessions, product views, carts, checkouts, and purchases; GA4’s funnel exploration will draw the same picture with more knobs. Pull the last 30 days, compute each step as a percentage of the step before it, and circle the ugliest ratio. Then split the same funnel by device, because a healthy blended number routinely hides a broken mobile step.

Reading the circle

• Weak session-to-product-view: a routing problem. Home page, navigation, collection pages. The traffic came; the store failed to show it anything.
• Weak view-to-cart: a persuasion problem. Product page photos, price framing, fit doubts, missing proof.
• Weak cart-to-checkout: a commitment problem. Shipping surprises, forced hesitation, a cart that hides the exit to payment.
• Weak checkout-to-purchase: a friction problem. Payment options, errors, forms. The rarest leak on Shopify, and the most urgent when real.

Write down your worst step and its ratio next to your tip 1 baseline. From here on, when a tip in this series targets that step, it jumps the queue. That one sentence of triage will save you a month of polishing pages that were never the problem.

Conversion rate, average order value, and revenue per visitor: baseline these three Shopify metrics this week, before you change a single thing on your store. That is the entire tip. It is also the one most stores skip, which is why most CRO effort is indistinguishable from luck.

Here is the uncomfortable truth about conversion advice, including the 99 tips that follow this one: if you have no baseline, you cannot tell whether any change worked. You will ship a new hero image the same week an influencer mentions you, watch sales tick up, and credit the hero image. The store learns nothing. You just feel productive.

The three numbers

Conversion rate (CVR) is the percentage of sessions that end in a purchase. Average order value (AOV) is what a purchase is worth. Multiply them and you get the number we would keep if we could only keep one: revenue per visitor (RPV), what a single session is actually worth to your business.

Are these the only numbers a merchant should care about? Of course not. Your P&L runs on margin, acquisition cost, repeat rate, returns, lifetime value. But those are moved by ads, pricing, and retention. These three are the scoreboard for the store itself: they are what moves when your website gets better at its one job, turning a visitor into an order. That is the job the next 99 tips attack, so this is the instrument panel we will fly by. Supporting gauges like add-to-cart rate and funnel step-offs join the panel in later tips.

Where to get them

All three live in your Shopify admin under Analytics. Set the date range to the last 30 days, note conversion rate and average order value from the dashboard, and divide total sales by total sessions for revenue per visitor. Ten minutes, no apps, no spreadsheet gymnastics required.

Date the numbers when you write them down. A baseline without a date is trivia. And use 30 days, not 7: one weekend sale or one slow Tuesday will lie to you in a weekly window.

A word about benchmarks

The first thing everyone does with a fresh baseline is Google whether it is “good.” Resist. Published averages like Littledata’s conversion benchmarks are fine for orientation, but a $30 impulse-buy store and a $900 furniture store should never be graded on the same curve. Traffic mix alone can double or halve your conversion rate without you touching anything.

The only benchmark that matters for the next 100 days is your store, 30 days ago. Beat that, repeatedly, and Cyber Monday takes care of itself: every point of improvement you bank now gets multiplied by the biggest traffic of your year.

That is the whole game, and it starts with three numbers on a sticky note. Go write them down.