YT YouTube2Text

YouTube2Text → Transcripts → Copy This CRO Tutorial, It'll Blow Up Your Shopify Store

Copy This CRO Tutorial, It'll Blow Up Your Shopify Store — Transcript

by Arsh Sanwarwala | Ecom Landing Pages + CRO · 7,152 words · 1,033 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00I have easily audited over 1,000 Shopify
0:02stores. Some came to me looking
0:03beautiful, while others came to me
0:05looking genuinely rough. But almost all
0:07of them had the exact same problem,
0:09which is that they were leaking money.
0:10So, in the next few minutes, I'm going
0:11to walk through the exact conversion
0:13checklist that I run for my client
0:15stores, section by section, step by
0:17step, so that you can copy it for your
0:19own store today. Here's the one thing
0:20that I really want you to get. Most
0:22store owners and Shopify brands focus
0:23all of their effort and energy into
0:25making their websites look good, but
0:27looking good is not the same as making
0:28money. A store can look gorgeous, but it
0:30can still convert terribly. The part
0:32that actually turns the traffic that
0:33you're already paying for into sales,
0:35the conversion engineering underneath
0:37the design, that's the part that almost
0:38nobody does. And it matters more now
0:40today than it ever used to be, because
0:42acquiring a customer is 60% more
0:44expensive today than it was 5 years ago.
0:46The traffic you already have is the most
0:48expensive it has ever been. Pour that
0:50into a store that leaks, and you're just
0:52paying more to lose more. So, stick with
0:54me, because each of these steps plugs a
0:55specific leak, and the last one is the
0:57one that separates a pretty store from a
0:59store that actually makes money. Okay,
1:00so before we get into the actual steps,
1:02it's very, very important to understand
1:04what conversion rate optimization, or
1:06CRO, actually is. Now, most business
1:09owners, most Shopify brands, and even
1:10most agencies, right, they think that
1:12CRO is things like changing button
1:14colors, following AB testing best
1:16practices, right, you often see this
1:17perpetuated a lot, or copying the
1:19biggest competitor, or even just copying
1:21what other big brands are doing. And
1:22that last one is always the most funny
1:23to me, because, and I get this all the
1:25time, right, Shopify brands come to us
1:26and they say, you know, I copied this
1:28competitor, I or, you know, I noticed
1:30that this big brand is doing this, and I
1:31tried to replicate that, and their
1:32stores still are not converting super
1:34profitably or converting where they want
1:35it to be. And the obvious logic is, if
1:38everyone was able to copy their top
1:39competitor, or copy what the top brands,
1:41the highest converting brands are doing,
1:43literally everyone would be converting
1:44at 10% and be making millions and
1:46millions of dollars, right? But
1:47obviously, that's not the reality, and
1:49that's not the case. And so, the
1:50important realization is that CRO,
1:52conversion optimization, is not about
1:54testing best practices or button colors
1:56and all this trivial stuff. What
1:57conversion rate optimization actually
1:59is, right? And what you must realize is
2:01that it's about understanding your
2:03customer at the deepest level possible
2:05and it's about collecting data, right?
2:07And making data-backed decisions. Your
2:09data will become the hypothesis for
2:11every single change you make. Okay, so
2:13there's two different types of data that
2:14you need. There's quantitative data and
2:15there's qualitative data. So,
2:17quantitative data is the actual numbers,
2:19right? It's things like heat map data,
2:21so where people clicking, where they
2:22scrolling, um how far do they scroll,
2:24looking at Google Analytics metrics like
2:26time on page, bounce rate, conversion
2:28rate, all of these different metrics,
2:29right? So, quantitative data tells you
2:31what is happening. Qualitative data, on
2:33the other hand, is opinions, feelings,
2:35and reasons behind those numbers. So,
2:37typically qualitative data comes in the
2:39form of things like interviews, uh
2:41conducting surveys, conducting review
2:42mining, where you kind of, you know,
2:44look at all your reviews and um analyze
2:46the sentiment. It's all of the kind of
2:47qualitative things like the feedback,
2:48right? The verbal things that people
2:50say. So, qualitative data tells you why
2:52it's happening. And what we do
2:53differently at our agency is we
2:56over-index super hard on the qualitative
2:58side, right? So, the things like
3:00interviews and surveys and review mining
3:01because that historically has been where
3:03we have found the biggest conversion
3:04lifts ever. It's just had an outsized
3:06return to relative to the quantitative
3:08data, which is kind of table stakes.
3:09Everyone is doing the quantitative, but
3:11very few people really, really
3:12over-index and really focus on the
3:14qualitative data. Okay, so let's walk
3:15through the most common types of
3:17qualitative data and I'll walk through
3:18exactly how you can collect it cuz I I
3:19want to make this super practical. So,
3:21the first form, classic form, is
3:23customer interviews, right? This one is
3:25super old school, but almost nobody does
3:27it. Literally just pick up the phone and
3:30call your customers. So, what we do for
3:31our clients is we will email their
3:33customers offering them some sort of
3:35incentive like, you know, $50 gift card,
3:37$100 gift card, whatever it may be, in
3:38exchange for 1 hour of the customer's
3:40time. And what we do is we vet them
3:42through a Google Form, right? so we'll
3:43typically have a couple different
3:44segments. We'll vet people based on are
3:47they recent buyers, so they've just
3:48recently purchased a product, are they
3:49long-time repeat customers, right? We'll
3:51kind of segment them out in different
3:52categories because, you know, they kind
3:54of think very differently and we want to
3:56hear both of their perspectives. And
3:57what we'll do is we'll basically
3:59interview the customer. It's almost like
4:01a real skill in how you run the call
4:03because you don't want the interview to
4:04come across as if you're actually
4:06interviewing them, right? Cuz if you
4:07come across as if you're interviewing
4:08them, the customer is just going to give
4:10you scripted useless answers, right?
4:11It's not actually going to go deep
4:13psychologically into how they're
4:14thinking about, you know, their daily
4:16life and how their product fits into
4:18their daily life. So, there's an art to
4:19conducting the call and you don't want
4:21it to feel like an interview. So, here's
4:22a template that you can follow, right?
4:24I'll put a link in the description below
4:25for this, which is basically kind of
4:26like an interview framework, right? So,
4:28typically the way that we start off
4:29these calls is we just set the frame. We
4:31make the person feel like they're not
4:32being tested or interviewed formally,
4:34right? So, we'll say, you know, this is
4:35casual. We're not looking for perfect
4:37answers. There's no wrong answers at
4:38all. We just want to understand, you
4:40know, what your life looked around the
4:41time that you started looking for a
4:42solution. So, feel free to answer
4:43naturally. Um, you know, there's no
4:45wrong answers, right? We kind of set the
4:46frame in that form. Uh, we typically do
4:48a bit of icebreakers as well. So, we'll
4:49kind of just set the scene, you know,
4:51ask them for a fun fact. Just kind of
4:53break the ice cuz again, you're trying
4:54to get their guard to go down. You
4:55really want to get authentic real
4:57answers. And again, this is why there's
4:58an art to this because the tonality of
5:01the way you ask the questions is super,
5:02super important. And then for the
5:03interview itself, right? You're just
5:05asking really simple questions. So,
5:06first you want to understand their life,
5:08right? So, you're not mentioning
5:09anything about the product up until this
5:10point. In the beginning, you're just
5:12asking, "What does a normal week look
5:13like for you? Walk me through a regular
5:15day from the moment you wake up to the
5:17moment you sleep. What does your
5:18day-to-day routine look like, right? And
5:20what do you typically do in that
5:21day-to-day? What takes up your time,
5:22energy, or attention?" Then we ask them
5:24questions about, you know, what parts of
5:25their week feels rushed, right? So, what
5:26are common pain points throughout the
5:28week? Um, what's something you put up
5:29with even though it was bothering you?
5:30And was that something that happened
5:31once in a while or was it becoming a
5:33regular thing? So, we're basically
5:34gradually building up to the product,
5:36right? We're trying to get the full
5:36picture of who they are as a person and
5:38what's going on in their day-to-day
5:40life. And then eventually and gradually,
5:42we kind of lead into the solution and
5:43talking about the product. So, then
5:45we're kind of segueing into how did
5:46they, you know, find out about our
5:48product? You know, what was going on in
5:49their life that even made them think
5:51about this product? You know, finding
5:52the triggering moment, um understanding
5:54how they search for it, understanding
5:55any doubt or decision criteria, like
5:57what other alternatives were they
5:58considering? You know, this is just a
5:59rough framework. You don't have to
6:00follow this step-by-step. Quite frankly,
6:02we don't even follow this step-by-step.
6:03Like we just go with the flow. Towards
6:05the beginning of the interview, we'll
6:06kind of keep it a bit more structured
6:07and rigid, kind of hitting on those key
6:09points that I mentioned earlier. But
6:10then, we just kind of go with the flow,
6:12right? So, if they say something
6:13interesting, you know, um you can kind
6:15of follow up and probe deeper. Just be
6:16genuinely curious, right? Ultimately,
6:18you know, this is a customer with real
6:19emotions and real feelings, and you just
6:21want to tap into that and really
6:22understand what was going on in their
6:24mind, right? When they were buying your
6:25product. So, this is an absolute
6:27absolute goldmine for our larger
6:29clients, eight to nine-figure Shopify
6:31stores, we'll often do like 10 to 20 of
6:33these in just the first month. And these
6:34are like 1-hour long each. Some of them
6:36even go like two to two and a half hours
6:37long. Again, it's very labor-intensive,
6:39very time-intensive, especially in the
6:41world of AI right now, where everyone's
6:42looking at all these different tools,
6:44and, you know, these apps to increase
6:45conversions, people forgot about the
6:47fundamental basics, which is understand
6:49your customer, right? The old-school
6:50way, just understand your customer, and
6:53they will give you the answers. Sounds
6:54simple, very few people do it. Okay, so
6:55the second form of qualitative research
6:57is what I call user testing. Unlike the
6:59user interviews, which we just
7:00discussed, those are live, right? Those
7:01are like one-on-one with a person on a
7:03Zoom call or Google Meets. With the user
7:05testing, you're actually testing your
7:06website with your users. Whereas with
7:08the interviews, it's just an interview.
7:10It's you're not actually testing the
7:11website. So, for user testing, what
7:13you're doing is you're getting people to
7:14go through your website that fit within
7:16your exact target demographic, right?
7:18And you can filter through that with
7:19screening questions, and then you're
7:20asking them to go through a series of
7:22different tasks on your website. So,
7:23what were your first impressions upon
7:24landing, right? Can you go through the
7:26process of uh finding this product on
7:28your store and talking aloud through
7:29your experience? So, we use a tool
7:31called UserBrain for this. Costs around
7:32$45 per test, so it's relatively
7:34inexpensive. And these are examples of
7:36some user tests we ran for our apparel
7:38client here, right? So, if I click into
7:39this, right? This is a 20-minute long
7:41recording of this person, right, this
7:43customer going through our client's
7:45website and answering a series of
7:46different questions. You can see over
7:47here that we actually screen them. So,
7:49our apparel client is selling hats, and
7:51so we only wanted to filter for people
7:53that own a lot of hats, right? So, we
7:54asked them, "How many hats do you
7:56currently own?" And we're only accepting
7:58people that say 20 or more hats. If they
8:00say any of these other options, we
8:01simply just reject them from the test.
8:03So, this way you're only filtering for
8:04your exact target demo. And so again, if
8:06I actually play this recording, right,
8:07this is just them going through the
8:08website answering questions. Right, and
8:10we set up all of these here. Now, if you
8:12want a framework for what questions to
8:13ask, again, I have this uh template over
8:15here. I'll put a link in the description
8:16below. Um this kind of goes through like
8:17the screening questions, which of course
8:18will be, you know, subject to your
8:20specific brand and what's, you know,
8:21relevant for you. And then you can have
8:22your screening questions here, right?
8:24So, you can choose whether to approve or
8:25reject them. Um and then you have your
8:27main testing questions. You can kind of
8:29follow this framework and just
8:30substitute your company name here for
8:31your name. But fundamentally, we're
8:33trying to get all of their qualitative
8:35feedback. And this is so crucial. Again,
8:37a heat map, quantitative data, can show
8:39you that everyone clicks on one element
8:41on your page, right? You can tell that
8:42with a heat map. But it's never going to
8:44tell you why they click on that element.
8:46A 25-minute recording of your ideal
8:48customer, your exact target demographic,
8:50getting confused, getting frustrated,
8:52right, is how you finally get that why.
8:53Through these user interviews, we're
8:54able to uncover what people think of the
8:56website, where do they get stuck, what's
8:57their level of confusion around certain
8:58elements, what kind of words are they
9:00using to describe certain things. It's
9:02just a gold mine of data. Okay, the
9:03third piece of qualitative data that you
9:05can collect is surveys. Now, this is one
9:06of those things where you need to have a
9:08lot of traffic in order to actually make
9:09this effective, cuz if you only get like
9:11a thousand visitors a month or
9:12something, you know, uh it's not a ton
9:14of data to really do the surveys. There
9:15is typically two types of surveys we
9:16run. The first type of survey is a
9:18post-purchase survey, right? So, this
9:19fires right after somebody buys. Uh we
9:22use a tool for this called Okendo. Okay,
9:24so just as an example here for one of
9:25our clients, we set up a post-purchase
9:27survey saying, um you know, "What
9:29motivated you to purchase from us
9:30today?" Right, and we kind of set up
9:31them some preliminary multiple-choice
9:33options, right? So, you can see these
9:34were kind of the responses here. Then we
9:36asked them, "What one factor influenced
9:38your decision to purchase the most,
9:39right? And again, we gave them some, you
9:41know, close-ended answers, multiple
9:42choice. And then this one is the biggest
9:44gold mine, which is, "Did you have any
9:46concerns, doubts, or uncertainties that
9:48almost stopped you from purchasing
9:49today? If so, what were they?" So,
9:51they've already purchased the product,
9:52so they've ultimately expressed
9:54interest, but what kind of stopped you
9:55from buying? You can see for this
9:56client, right, they sell gas masks. So,
9:59on the right-hand side are all the
10:00quotes from people, right? Some people
10:01said, "More information on fit." Forum
10:04sites like Reddit, "Say your mask run
10:05big and don't fit well." Boom. Just got
10:07a super golden piece of insight. And now
10:09I can think like, "Okay, I want to see
10:10how can I alleviate this concern on the
10:11website?" The second person said, "More
10:13layman terms for filters, right, for
10:14their gas masks. Just want to make sure
10:16I'm purchasing the right ones." The
10:17third one, "My main concern is if the
10:19mask will fit my face or head properly."
10:21So, again, something related to fit.
10:22Fourth one, "Sizing concerns." The fifth
10:24one, "Looking for a discount code but
10:25couldn't find it." Right? So, you're
10:26collecting so much gold mine of data
10:28here that's incredibly valuable. And
10:30then another question we asked was, "If
10:32you were considering another brand, who
10:34were they and what led you to choosing
10:35us instead?" Right? So, then this gives
10:37you a ton of insight into what
10:38alternative solutions people were
10:40considering to kind of really put you in
10:42the mind of the customer. So, that was a
10:43post-purchase survey. Now, there's
10:44another survey that we also run called
10:46the product friction survey. The way
10:47that this works is, let's say your
10:48average customer spends a minute on your
10:50website, right? That's the average time.
10:52What you're going to do is you're going
10:53to trigger a pop-up around the 70% mark
10:55of your average time on page and ask
10:57them, "What is stopping you from
10:59buying?" Right? So, this is how we
11:00structured it for a client. "Is there
11:01anything holding you back from buying
11:02this product right now? Yes or no?" If
11:04they say yes, which was 61%, we say,
11:06"What's the reason that's holding you
11:07back from buying?" And then we leave it
11:08open-ended, right? And then you can see
11:10on the right, these are all different
11:11things that people are saying, very
11:12specific stuff that we would have never
11:14known had we not run this survey. Okay,
11:16the final thing is what we call review
11:17mining. So, basically, what you're going
11:18to do is you're going to pull all of
11:19your reviews from Amazon, Trustpilot,
11:21your website, wherever they live into a
11:23spreadsheet. And you're going to use my
11:25prompt via ChatGPT or Claude to pull the
11:27most common words, topics, and themes,
11:29right? So, you take that exact language
11:30and inject it back into your copy. The
11:31best copy on your site comes straight
11:33from the words of the customers that are
11:35already buying your product. So, these
11:36are all the reviews for our client here,
11:38right? They have thousands and thousands
11:39of reviews. So, we put it all together
11:41in a spreadsheet, and all we simply did
11:43was throw the reviews into ChatGPT and
11:45threw in a specific prompt, which I'll
11:46show you in a second. What the prompt
11:47does is it basically pulls the most
11:49frequently mentioned words, right? So,
11:50what's the most commonly mentioned words
11:51that people are saying? As you can see
11:53here, around uh for a gas mask product,
11:55um they're using words like filter,
11:56filtration, right? Comfort,
11:57comfortability, price, money, value,
11:59eyes protection, etc., right? And it's
12:00mentioning how many times they mentioned
12:02it. Uh I actually pulled a smaller
12:03sample size just for this video, but
12:05they actually have thousands of reviews.
12:06What the prompt also does is it pulls
12:07the most frequently mentioned phrases,
12:09right? So, um what are the most commonly
12:11mentioned topics that people are saying?
12:12This is all the data from people that
12:14have already bought, and all you're
12:15going to do is you're going to inject
12:16this back into your copy, which I'm
12:17going to walk through exactly how you
12:18can do that later. This is a prompt over
12:20here. So, again, I'll throw a link in
12:21the description below to access this.
12:22But, you just throw in your reviews, and
12:23you just throw this into ChatGPT or
12:24Claude, and you'll get some uh good
12:26feedback here. Now, the next part is the
12:27quantitative data, right? So, this is
12:29the actual numbers. The first tool that
12:30we use is called Microsoft Clarity. This
12:32is a free tool, which shows you where
12:33people click, where they scroll, and you
12:35can actually watch recordings of people
12:37on your site. Okay, so, I have a client
12:38account pulled up here just for
12:39demonstration. So, in the last 30 days,
12:40they got 226,000 visitors. And if we go
12:43into the heat map data over here, we can
12:45actually look at the breakdown per page.
12:46So, if I just click on this over here,
12:48we're actually able to pull their click
12:50map data. So, as you can see in this
12:51tool, right? I'm actually able to see uh
12:53where all of the clicks are going on
12:55different elements, right? So, I can see
12:562,100 clicks went here. Uh 2,800 clicks
12:59went here. And I can actually filter by
13:00area on the left, and I can see the
13:02percentage of clicks that go to
13:03different elements. So, if I click edit
13:04area, right? I can click on different
13:05blocks, and I can see the clicks. And
13:07even you can see how far people scroll.
13:09So, I can see 77% of people have
13:11scrolled this far, right? 56% of people
13:12have scrolled this far. The way that I
13:13like to think of it is are people
13:15clicking where I want them to click? Are
13:17people scrolling where I want them to
13:19scroll? And if they're not clicking
13:20where I want them to click, or they're
13:21not scrolling where I want them to
13:22scroll, how can I reorganize the
13:24content, you know, change things up,
13:26change the copy, the design, the
13:27structure, etc. Such that I'm able to
13:30get them where I want them to click and
13:31scroll where I want them to scroll,
13:32right? It's as simple as that. Okay,
13:34quantitative metric number two is Google
13:35Analytics. Now, there's a couple of main
13:36reports that I like to pull here. Okay,
13:38so the first thing I like to go and do
13:39in Google Analytics is I like to go into
13:41reports and on the left-hand side,
13:43you're going to open up uh drive sales,
13:45right? And then you're going to look for
13:47something called the purchase journey
13:48report. So, inside of the purchase
13:49journey report, you're going to filter
13:50for let's just say like the last, you
13:52know, whatever time period, last 90 days
13:53for example, right? That's a good sample
13:55size. Um and what you're going to see
13:56over here is basically the full funnel.
13:58So, from the moment somebody starts on
14:00the website, right? Session start 100%,
14:02how many people are viewing the product,
14:04how many people are adding to cart, how
14:06many people are beginning checkout, and
14:07out of that, how many people are
14:09actually purchasing. Now, this is a very
14:11simple framework that I like to follow,
14:13which is called the metric on fire. So,
14:14every company has one metric, right? One
14:17metric on their store, where they're
14:19under indexed or underperforming on, aka
14:21the metric on fire. And your goal is to
14:23identify that metric on fire and focus
14:25all of your energy and effort on fixing
14:27that metric on fire in order to increase
14:29your conversion rate. Because with CRO,
14:31there's so many different things you can
14:32test, but again, coming back to what I
14:34said in the beginning, it's about making
14:35data-backed hypotheses. So, use the data
14:38to inform your insights. So, where are
14:39we under indexed on this store? So, the
14:41view product rate right now is 40.3%.
14:43The target benchmark for this should be
14:4550 to 70%, right? So, we're slightly
14:48under indexed here on the lower end.
14:49We're at 40, whereas we should be at 50
14:51at a bare minimum. So, slightly under
14:53indexed there, but not super
14:54underperforming. In the 1,000-plus
14:55Shopify stores that I've audited, I can
14:57tell you without a shadow of a doubt,
14:59the view product rate is in 90% of the
15:01cases the most underperforming or most
15:02commonly underperforming metric, right?
15:04So, for most Shopify brands that I audit
15:06and for many of you watching here, it's
15:08very likely that your view product rate
15:09is underperforming and that's your
15:11metric on fire, right? That's by far the
15:12most common. Then you have the add to
15:14cart. So, for this client, it's 22.7%.
15:17The benchmark for this is 12 to 20%,
15:19right? So, again, over here, we're
15:21actually well in the clear. Like, our
15:22add to cart rate is not underperforming.
15:24It's not underindexed. But, this is
15:26usually for most clients that I audit,
15:27usually the second most common metric on
15:29fire. So, the first is the view product
15:30rate, the second is the add to cart. The
15:32next metric is the begin checkout rate.
15:34So, for these guys, it's 33.4%. The
15:36benchmark for this should be 40 to 60%.
15:38Again, we're underindexed over here. And
15:40then the final metric you have is a
15:41purchase rate, right? So, the purchase
15:43rate is 64%. The target benchmark for
15:44this is 45 to 60%. So, again, we're well
15:47above KPI over here. So, realistically,
15:48the two main metrics on fire over here
15:50are the view rate followed by the begin
15:52checkout. The second report that I like
15:53to pull is a landing page report. So,
15:55you're going to go into reports over
15:56here, and you're going to click on
15:57landing page, and I'm going to scroll
15:58down here, and I'm going to add a couple
16:00key metrics to this report. So, I'll
16:01click on metrics. I'm going to add the
16:03average purchase revenue per active
16:05user, aka the revenue per user. So, how
16:07much revenue are you making per user?
16:08You're going to pull up the bounce rate,
16:09and that's pretty much it, right? Bounce
16:10rate, revenue per session, and average
16:12engagement time per session. So, click
16:13apply. And now we can see this is all
16:15the data. So, what I try to do here is I
16:17try to look for anomalies in the data,
16:18right? So, this is a landing page
16:20report. So, basically, it's measuring
16:21from the moment somebody lands on the
16:23website, like all of the different pages
16:24that they can land on, what's the
16:26conversion rate for each of those pages,
16:27what's the bounce rate, so how many
16:28people are exiting from the site, what's
16:30the average engagement time, all of
16:31these different things. So, one really,
16:33really simple, practical insight that we
16:35noticed was the homepage over here is
16:37converting at 2.37%, right? This was a
16:39while back when we were doing testing
16:41for them. Now, it's much higher. And
16:42their collection page converts at 1.57%,
16:44okay? So, they were actually driving all
16:48of their paid traffic to the collection
16:49page, even though their homepage
16:51converts at 2.37%.
16:53Also worth noting is that the revenue
16:55per user on the homepage is 1.67. The
16:57revenue per user on the collection page
16:58was 1.10, right? The bounce rate was
17:00lower on the homepage. The average
17:02engagement time was higher on the
17:03homepage. So, effectively, every single
17:05metric across the board was higher on
17:06the homepage. The client didn't even
17:08know this, and they were driving their
17:09traffic to the collection page. And so,
17:10one of the first tests we ran for them
17:11was simply just sending traffic to the
17:13homepage. And all the data pointed to
17:15this, so it was an obvious no-brainer.
17:16So, that's a real-life practical example
17:18of how you can use a landing page
17:19report. All right. So, now we've covered
17:21both the qualitative and quantitative
17:22data. The next step is putting it all
17:24together into an actual plan of action
17:26and what we're going to test on our
17:27store. So, before you run a single test,
17:28you need to know which type of CRO
17:30you're even allowed to do cuz it comes
17:31entirely down to your traffic or order
17:33volume. Scenario one is if you're under
17:3550k to 100k visitors a month. Let's just
17:37call it 50k, right? And you're fewer
17:39than 400 conversions, you cannot run AB
17:41tests because you don't have enough data
17:43to do so, right? There's something
17:44called statistical significance, which
17:45is whether your test is due to chance or
17:47not. So, if you don't have enough data,
17:49you can't run AB tests and you have to
17:50go for bigger swing changes like a new
17:52landing page, a large-scale rebuild, and
17:54whole new section. So, you can still
17:55make data-backed improvements, you just
17:57cannot do it in the form of an AB test.
17:59If you are above that threshold, you're
18:00doing 400 plus orders a month and 50 to
18:0270k visitors, again, just call it 50k
18:04here, you can AB test, right, to
18:06significance. But this is like a bare
18:07minimum. So, when you have a lot of
18:08data, you can test small things like a
18:10headline, the above the fold, an image,
18:12you can get away with those changes,
18:13right? And you actually do want to do it
18:14this way. You don't want to go for big
18:16swings cuz that can literally tank a
18:17store that's already performing.
18:18Whereas, when you don't have much data,
18:19you have to go for those bigger swings
18:21because you don't have enough data. But
18:23again, you're doing all of the changes
18:25based on actual data that you collected
18:26based on what I covered earlier. Okay,
18:28so I'm going to walk through all of the
18:29tests and changes that I would make and
18:31give you a couple examples and I'm going
18:32to show you the data behind it. So, I'm
18:34actually going to show you all of the
18:36various data, the qualitative
18:37quantitative data, and the before and
18:38after result, right? I'm doing it this
18:40way to prove that when we increase
18:42conversions, these are not random
18:43changes because every single one traces
18:46back to something the data told us and
18:48that's why I spent the start of this
18:49video just focusing on data collection.
18:50Okay, so one of the first things you
18:51want to do, again, when you start making
18:52tests, is centering it around the metric
18:54on fire. So, we're going to start with
18:55the most commonly mentioned metric on
18:57fire, which is the view product rate. If
18:59Basically, the goal here is to increase
19:00the percentage of people that view a
19:01product on your store and by doing that,
19:03you're going to increase all your other
19:05metrics, your add to cart, your begin
19:06checkout, and your purchase and
19:07conversion rate. So, in order to
19:08increase the view product rate, the
19:09first place you want to start is the
19:11homepage. It's the most viewed page,
19:13right? Because um a lot of people who
19:15send paid traffic, they send it to a
19:16landing page or a product page, but a
19:18lot of people will still go and Google
19:19your company. They'll, you know, come
19:21across an organic piece of social media.
19:22Somehow they land back on your home page
19:24or even they'll click from the ad back
19:25to the home page, right? So, if you look
19:27at your GA4 data, you're going to find
19:29your home page one of the most visited
19:30pages. Okay, so for our pet e-commerce
19:32client, we kind of created this audit
19:34report for them and part of the audit
19:35report, one thing that we learned from
19:37the data, where you see the number two
19:38symbol over there, you can see that the
19:40one of the most clicked elements, the
19:41seventh most clicked element, was this
19:42view all button on the best seller
19:44section, right? Um and right underneath
19:46that is the product images. Basically,
19:48more people were clicking on the view
19:50all button to view all the products
19:52instead of the individual products,
19:54which naturally makes sense because
19:55they're selling, you know, clothing
19:56apparel for dogs, right? So, people want
19:58to see all of the designs. They don't
19:59want to click into individual ones. And
20:01when you scroll down further on the home
20:02page, when you looked at their bundle
20:04options, again, a majority of the clicks
20:06were going towards the view all button.
20:08So, just to show you an action here,
20:09right? When you scroll down, you can see
20:10this view all button. When we looked at
20:12that Microsoft Clarity heat map that I
20:13just showed you, majority people were
20:14clicking on the view all rather than the
20:16products, right? And as you scroll down,
20:18they clicked on view all over here and
20:19then if you scroll down more, they
20:20clicked on view all over here, view all
20:21over here. So, all of the clicks were
20:23going to the view all button for
20:25basically directing people to the
20:26collection pages. However, despite being
20:28one of the most clicked elements on the
20:29page, when we looked at the scroll map
20:31data, a vast majority of people were not
20:33even reaching these sections in the
20:34first place, right? because it's so
20:36further down on the page. And so, a
20:37really simple AB test that we ran based
20:39on the actual data was we took that heat
20:41map data where people were clicking and
20:43we basically moved it up to the top in
20:44the form of these social media icons as
20:46you can see here above the fold that
20:47basically um showed a really cute photo
20:49of a dog wearing the product, but when
20:50you clicked into one of these items, it
20:52would take you into the collection page.
20:53And so again, this was a data back
20:55hypothesis because we already validated
20:57that people were looking for this. They
20:58were clicking on it further down on the
20:59page, but very few people were even
21:01reaching the section. So, we moved it
21:02up. And again, what we look for is
21:04multiple data points. So, it wasn't just
21:05the quantitative data telling us this.
21:07It was also the fact that people
21:08mentioned in user interviews that they
21:10would love to be able to get into the
21:11products much quicker. They were
21:12actually showing frustration when we
21:14were doing the user testing saying that
21:15they couldn't find the right products.
21:16They had to look for it, right? So, we
21:18want to make it super easy. And this
21:19test absolutely crushed it for them. We
21:21saw 72% lift in their conversion rate
21:23and we drove an incremental 83% increase
21:26in the revenue per visitor just as a
21:27result of this AB test. Okay, the next
21:29area to focus on to increase your view
21:30product rate is your navigation. Now,
21:31again, you can use multiple data points
21:33to support this insight. So, for our pet
21:35e-commerce client, we looked at their
21:36navigation menu and we pulled up
21:38Microsoft Clarity, right? So, this was
21:39their before version. So, you can see
21:40here that the shop button got 0.7% of
21:43clicks, the info button got 0.05, there
21:46was a tab called just for fun 0.12% of
21:48clicks, and our story got 0.05, right?
21:50And then all of these other items like
21:52login, create account, search measly
21:54like close to 0% of clicks. The other
21:56thing that people said in the user
21:57interviews that we conducted, so this is
21:58the qualitative data side, was that
22:00people couldn't find the right product.
22:01They found it very confusing, right? So,
22:03we had multiple supporting data points
22:04that say that, "Look, this navigation
22:06based on the data needs to be redone. We
22:07just want to get people into shop as
22:09quickly as possible." So, what did we
22:11do? You recall this test we ran over
22:12here with the social media icons which
22:14drove a 72% lift in conversions. We took
22:16that same premise of that test and we
22:18completely cleaned up the navigation. We
22:20stripped it down to the most important
22:21items which are just shop. Now, I don't
22:22have to click into shop. Before I had to
22:24click into shop. We just got rid of all
22:26of that. This was a site-wide test,
22:27right? Because it's on the navigation
22:29and we saw a 15% lift in their
22:30conversion rate from just this one test.
22:32Okay, so the final area to focus on to
22:33increase your view product rate is the
22:35collection page. And so, there's a
22:36couple different things you can do here.
22:37Again, I'm going to walk through actual
22:38test that we ran. So, first thing that
22:39we did for our client when we pulled up
22:42their data in Microsoft Clarity on the
22:44product page, we noticed that this guy
22:45right over here, the little dots, this
22:47is basically their product thumbnails,
22:48right? When you go to product page, you
22:49have the thumbnails of the products you
22:50can scroll through. And this got 8.89%
22:52of clicks. When we looked at Google
22:54Analytics, we determined that the
22:55percentage of people that bought a
22:57product after clicking on one of these
22:59images was exponentially higher than any
23:01other element on the store. And so, we
23:03wanted to think, "How can we leverage
23:05this data across other areas of the
23:07website?" And so, we did exactly that.
23:08The goal of the collection page is to
23:09help people find the right product. And
23:11so, because we had the data to support
23:13it, both qualitative and quantitative,
23:14we added in the product image thumbnails
23:16on the collection page itself. So,
23:18people could basically scroll through
23:20different angles of the product and see
23:21different images to basically help them
23:23find the right product in the first
23:24place. And from that AB test, we saw a
23:2622.2% lift in their conversions and we
23:29saw a 7% lift in their average order
23:30value. We've also done this for numerous
23:32other clients, like our luxury rose
23:33client, where we saw the same data, 38%
23:35lift in conversions, 12% increase in
23:36their view product rate. Okay, so, up
23:38until this point, we've covered how to
23:39increase the view product rate, which
23:40again is the most common metric on Fire.
23:42The next area is the add to cart rate,
23:43right? This is the second most common
23:44metric on Fire. Now, increasing the add
23:46to cart is directly correlated with
23:47improving your product page. The first
23:49thing that we did for a Gas Mask client
23:50is we noticed that this was driving a
23:52lot of the clicks, right? These little
23:53dots here that got 8.89% of clicks. This
23:56was one of the biggest determinators of
23:57whether somebody would purchase or not.
23:58However, the downside of this approach,
24:00and when we looked at the user testing
24:01data, was people said that this was
24:03really frustrating cuz they didn't even
24:04notice the small little dots on the
24:05store. And they said it was frustrating
24:07because if they had noticed it, it would
24:08have helped them to make a buying
24:09decision much faster. So, lo and behold,
24:11simple test was we changed the thumbnail
24:12dots to actual thumbnails, right? Like
24:14this, so that people could see what the
24:16visual of the product is and they're
24:17more likely to click into it. And that
24:18ultimately drove a 24% lift in their
24:20conversion rate. Another AB test that we
24:21ran for a pet client. So, in the before
24:23version on the left, right? You can see
24:25basically what they were able to fit
24:26above the fold, that the product image,
24:28the title, the the reviews, and three
24:30bullet points, right? And what we
24:31noticed in the heat map data was a lot
24:34of people were clicking on these
24:36thumbnails over here cuz they wanted to
24:37see the different styles and designs.
24:39And in the user interviews and the user
24:41testing that we conducted, people were
24:42saying that they would have loved to
24:43seen this much higher up on the page cuz
24:45that's what they were looking for. Like,
24:47they don't care about the details, they
24:48just wanted to see the designs and the
24:49styles cuz that's ultimately why they
24:51were buying. Based on that data, all we
24:52did was we moved the designs above the
24:54fold, so So can see in the after version
24:56on the right, it was above the fold
24:57immediately visible at the top, so
24:59people didn't have to scroll down for
25:00it. And what we saw was an increase in
25:02all metrics across the board. We saw 34%
25:04lift in the conversion rate, which again
25:05is not site wide, it's specific to
25:07product page to purchase, right? So that
25:09went up 34%. Revenue per visitor went up
25:1142%. AOV went up, add to cart rate went
25:14up. So basically everything went up. The
25:16final two metrics you have is your begin
25:17checkout rate and your purchase rate. To
25:19increase the begin checkout rate, you
25:20know, this will come down to things like
25:21trust signals in your checkout,
25:23you know, adding more things there like
25:25reducing fear and uncertainty like your
25:27money back guarantee and things like
25:28that in your checkout. Most clients are
25:29never really under index here. Since
25:31we've really optimized the store, we've
25:32managed to increase all these metrics,
25:33but you know, begin checkout is always
25:35one of those harder things. But for most
25:36people watching, it's not really
25:37something you need to worry about. And
25:39quite frankly, even for this client,
25:40we're not really focused on begin
25:41checkout, we're more focused on
25:42increasing the view product rate. And
25:43the final thing is the purchase rate,
25:44right? Which is 64%. This one we almost
25:47never touch because Shopify's checkout
25:48is pretty standardized, so you can't
25:50really do much over there and making my
25:51changes there. Okay, final thing just to
25:53end off the video, so everything up
25:54until this point was website level
25:56testing, right? The homepage, the
25:57product page, the collection page. The
25:58final piece of your CRO program is your
26:00landing pages. This is a separate video
26:02on its own, but the short version is
26:04that every improvement that I just
26:05walked through here can also apply
26:07there. You take the same copywriting
26:09principles, the same data-driven
26:10changes, the same structure, and you
26:12just apply them to your pages that
26:13you're sending your paid traffic to. The
26:14difference is that with your landing
26:16pages, you can make large-scale changes
26:17quicker and faster rather than small
26:19isolated AB test. And if you don't have
26:21data to AB test properly, right? And you
26:23want to do testing for your paid ads,
26:25all you have to do is just duplicate
26:27your ad campaign and point the traffic
26:29to a different landing page. So you're
26:31running both campaigns with real ad
26:32dollars behind them, and you're going to
26:34be able to see within a short window of
26:35a couple days which one is pulling a
26:37higher ROAS, right? If one is 5x ROAS,
26:39the other is a 2x, it's pretty obvious
26:40what the winner is. Running landing
26:42pages is pretty table stakes, you should
26:43be doing this regardless, but the thing
26:44that almost nobody does is the actual
26:46on-site AB testing, the website AB
26:48testing, which is why I try to focus
26:50really heavily on that, and that's where
26:51I think uh a lot of big brands really
26:53falter on cuz they focus so much on
26:55their paid traffic, they don't focus on
26:56their website, which is, you know, just
26:58leaving money on the table. Okay, to
26:59recap, collect qualitative and
27:00quantitative data first, find your
27:02metric on fire, test the highest impact
27:04areas, use the data that I walked
27:05through to inform your insights and make
27:07your changes, and then compound the
27:08winners. That is how you turn a Shopify
27:10store into a machine that compounds
27:11conversions every single month. So,
27:13that's a checklist step-by-step, and
27:15that's the exact one that I run on
27:16client stores. Run those yourself, all
27:18of them, right? And your store will
27:19probably probably convert better than it
27:20does today. And if you want to keep
27:22going, here's the next one to watch.

About this transcript

This page contains the full transcript of Copy This CRO Tutorial, It'll Blow Up Your Shopify Store by Arsh Sanwarwala | Ecom Landing Pages + CRO, generated from the public captions YouTube serves with the video. The transcript has 7,152 words across 1,033 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.