The amount of usable data that comes into a growing Shopify store is incredible. But too many store owners just sit on that data instead of putting it to use. Or even worse, they don't bother collecting it at all.
Collecting data is an important first step, but it's not enough. You must act on it. Being data-driven is all about humans collecting, interpreting, and acting on data.
The problem is that most store owners are already overwhelmed with running the business. Interpreting data and turning it into marketing and CRO decisions is a skillset you don't have years to learn.
You probably don't need to be collecting more data. You just need to know how to take the data you already collect, and turn it into insight and action.
What do you do with data? Make it part of your every day business processes
How do data and research lead to more profits for your store?
You are constantly faced with decisions about what you will do for your store.
- How do you get a higher ROI on your marketing dollars and efforts?
- What changes would make your site better at selling to your best customers?
- What problems cause you to leak revenue, and how do you fix them?
Without data to help you answer these questions, you are mostly going on guesswork, other people's experiences, or the marketing of people who want to sell you an app or service to fix a problem you may or may not really have.
Benefits of using data well
- Stop yourself from making bad choices.
- Verify the value of your investments.
- Avoid wasting time chasing down details that don't matter while something else tanks your conversion rate.
- Generally makes running your business easier and can make you BOATLOADS of money.
With data, you can better understand what problems you face and their root causes. You can know the thoughts and motivations of your customers. Most importantly, you can know that you are doing the right thing when you make decisions about marketing and what you need to make more sales.
Removing guesswork in favor of research-backed understanding gives you a better starting point for every move you make. You are less likely to make critical mistakes and more likely to make profit-boosting decisions. And either way, you will be measuring the impact of your actions, so you know for sure if the results are negative or positive.
Using data is literally like starting a mountain climb from the halfway point instead of the bottom.
Quantitative data & research are focused on numbers & quantifying critical issues surrounding your site.
- What are people doing on our site?
- How many people?
- How often?
- Where did those people come from?
- What parts of the site are working?
- What parts of the site aren't working?
You are looking to answer questions about what, who, when, and where. The main goal here is to find opportunities for improvement.
Forms of quantitative data
Examining the site by following a set of guidelines laid down by best practices and experience. This is mostly to create a starting point. We look at specific parts of the site and weigh them against things we know to be good or bad, without a need for testing.
No doubt, many of the tools you use have embedded analytics. Google Analytics for your site. Shopify Analytics for your store overall. Email marketing. Facebook ads. These give you numbers and (sometimes) visualizations about what is going on in each specific area.
Heatmaps visually show us what parts of the site are being seen and clicked on. So we know what is catching and holding the interest of would-be customers, and what isn't.
In addition to heatmaps, we can also record a user's session on the site. We can see the difference in how a buyer and non-buyer use the site, and how their experience might be different.
Real world application
You should go to your data with specific questions that you want answered:
- What traffic sources are turning into customers?
- What pages cause people to leave the site?
- What elements on a page do people have trouble using?
- Do we have issues with browser or device compatibility?
- What is the ROI on our marketing spend?
- Does a specific design element on the site help or hurt conversion?
- What is different about VIP customer behavior compared to lower-value customers?
The answers to these questions will help you understand what is happening on the website, and what needs to be fixed or optimized.
The information you get must be put into the correct context. You need to know what questions you are asking from your data, then go look in the right place for the answers.
That gives you a good starting point for understanding and better decision making.
- Hotjar- Heat & click maps, user recordings, form analytics. Hotjar is a really useful tool to visualize how people are using your website, and what stops them from becoming customers.
- Google Analytics- An insanely powerful & free tool for insights into your site and your customers.
- Google Tag Manager- Helpful for setting up marketing and tracking tags, as well as sending advanced event tracking through to GA.
- Google Data Studio- Use it to create effective dashboards from multiple data sources.
Qualitative data & research are about finding out why people do the things they do (or don't do). You want to see your site through your customer's eyes.
- Why do people do the things they do on your site?
- What are their motivations for buying or not buying?
- What are they thinking about when they are on your site?
- What specific role do you fill in your customers lives?
- What can you do to improve on the issues you find through quantitative methods.
Quantitative data uncovers what is going on. Before you can act on that, you need to understand the underlying issue that is causing people to act a certain way.
Forms of qualitative data
Using software to create on-screen popups, you can ask questions of your users. These questions should be few, short, and specific to the page they are looking at.
Surveys are on of the simplest forms of qualitative data mining. You can flat out ask your customers questions about who they are, how they think about your product, and why they buy from you. An email list segment of recent buyers is the best target for this kind of survey.
Recorded user testing
Recruit people to perform tasks on your site and see how it goes. These can be specific tasks like: "Find a pair of white, size 13 athletic shoes", or open ended: "Find a birthday gift for your spouse".
Examples of qualitative data use
- A survey leads shows that the store misunderstood exactly how customers used and viewed their products. A change to the value proposition message and product photography better suited the actual use of the product, and led to a 50% increase in revenue.
- Dropouts during checkout turned out to be caused by a phone number field. Buyers didn't want to give out their number for fear of marketing calls. Once the form was changed to explain that the phone number was only for potential delivery issues, completion rates shot up.
- A product that seemed fairly easy to use (by the people who created it) was something of a mystery to first-time buyers. It turned out that not knowing how exactly they would use the product (or how much they would need to use at a time) prevented them from buying. The addition of a few graphic instructions and a couple of short YouTube videos increased trust in the product and boosted conversion rates.
The answers you get through this kind of data are going to give you more insight into the issues you find & possibly uncover things you never thought of. This will help you make changes that are more likely to improve sales and site function right away, with less room for mistakes.
- A simple initial survey of recent customers. Long-past orders don't make good targets. This survey can be done in a few ways:
- Hotjar- Yes again. Hotjar has qualitative uses, mainly it's polling and survey functions. You can run on-site question pop-ups or longer-form surveys
- A data-collecting form. I use Paperform, but you could also use Typeform or another similar service.
- Email: You can send out your survey using your email service.
- User testing- You can recruit testers in person or through specific online services
- Ask your mom to try to place an order (or any friend/kin who aren't of above average technical ability)
- Get a drunk friend to use your site and see if they can buy from you
- Go into a coffee shop, buy gift cards, and then offer them to any patron that will try to accomplish a task on your site
You need accurate & dependable information to be effective at what you do. That doesn't mean you should have to dive into Google Analytics at the start of every day.
The numbers you need to check daily or weekly as part of your role within the business come from analytics. They are buried along with all of the other fine-grain details that can distract and confuse, when you should be working.
Data monitoring should fit in to your daily routine, and not eat up of all your time & attention. There should be a high-level view that you can just check with a glance.
That's where dashboards come in.
Dashboards as a decision-guiding tool
Dashboards contain high-level metrics that matter to you in your specific role(s) within the business:
- Owner/CEOs see high-level KPIs from multiple departments
- Marketers see KPIs about traffic sources and campaign ROI
- Financial people see revenue, profits, and expenditures
- Technical staff see site speed, errors, and evidence of anything broken
By creating dashboards that give you the information you need daily & NO MORE, you eliminate data overload and make data more useful.
Dashboards are most effective when they show exactly what the specific person looking at them needs to see on a regular basis, and nothing more. Operating this way allows you to use data to make sure things are going well, and allowing you to get back to your work.
Using dashboards does 2 things for you:
- They restrict the flow of information down to just what you need to see. This makes data into a decision making tool instead of a distraction and source of anxiety.
- They help you create a systematic use of data. You can set a time of the day or week to check out your dashboard. The key metrics you find there will tell you whether everything is humming along, or whether you need to stop and investigate a potential problem.
Data should be the main driver when it's time to seek out new ways to grow and optimize revenue for the business.
The most common thing to think about is conversion rate optimization (CRO), but other areas benefit from data in decision-making processes as well.
- Revenue Capture Strategy (retention, recurring revenue, average order value)
- Customer Service
- Product Innovation
Essentially, you are looking for a way to improve some specific part of your business, and using your data to make that possible.
This is when you go into your analytics reports with specific questions in mind, gain insight and answers from the data, and then act on it.
You might be asking questions like:
- Which marketing efforts are paying off, and which aren't?
- What pages on the site are leaking revenue, and how do we improve them?
- What problems are our customers running into when browsing the site?
- What demographics make up the highest spending 20% of my customers?
- What products are purchased together most often?
This is an intentional and process-driven use of data. You don't just stumble into analytics and find something that will increase your effectiveness. You go in and identify problems or new opportunities for growth via specific metrics.
Data-driven optimization process
By creating dashboards that give you the information you need daily & NO MORE, you eliminate data overload and make data more useful.
- Identify opportunity for improvement through analytics. (Quantitative data)
- A product page receives a lot of traffic, but the conversion rate is low.
- Research a solution for how to take advantage of that opportunity. (Qualitative Data)
- A user poll on the page informs you that visitors want to know what kind of material the product is made of.
- Create change to make an improvement (Design/Messaging Synthesis)
- Add the information that visitors need to a prominent place in the product description.
- Measure effect to see whether or not the change makes a difference, and whether that difference is positive or negative for the business. (Testing and Measuring)
-The product's conversion rate rises by 4%, indicating that many non-converting customers considered this information a priority, and now you make more product sales.
Remembering that you actually have analytics setup and strolling through random reports 4 times a year doesn't do any good.
Luckily, you already make all kinds of decisions for your business. You (or your team) have to decide what they will do each to day/week/month to get a positive ROI. You will always have to do that. Adding your data to the mix just means that you will make more profitable decisions and worry less about making the wrong choice.
Data should be used in service of revenue generation. Your business is made up of multiple interconnected systems that work together to effectively produce a profit.
To start out, think about those various systems that make up your business. They are a good point to start breaking down the various data sources and types you have, and deciding how you can use them.
With marketing, you need to spend your budgets of time, energy, and money where they will do the most good. How can you determine what makes for good marketing?
Since the goal of marketing is to drive high-quality traffic to the site, we judge marketing by how much traffic it drives and what those people do. 1,000,000 visits in a day would be amazing for most stores, but if none of them ever convert, it's all for nothing.
Use marketing data to see where you should spent your marketing budgets (time, energy, money) to best create purchase-ready traffic. So you have less waste, better effect, and more benefit for the business.
Sales (Conversion on the site)
The job of the website is to convert attention (visitors) into money (sales). The site does that through the story it tells visitors (messaging) and the way it looks and functions (design).
Data tells you how effective your messaging and design are at turning attention into money. It can break down that performance to the level of individual pages and even elements on those pages like category filters or Add-to-cart buttons.
Use conversion data to learn what does and doesn't work well for turning visitors into buyers. You can also use it to learn what changes to that design and messaging have the best chance of succeeding at converting more of them.
Customer service issues can be a goldmine of data for a store. Questions, requests, reviews, returns. Each and every one is a chance to learn exactly what customers want and expect from your store.
Data isn't just about numbers. Recording and using the thoughts and words of your customers is one of the most insanely valuable things you can do.
You can use the thoughts and words of your customers to see how you can give them the information and experience they need to make a better purchasing decision. That can help eliminate returns, chargebacks, and negative reviews before they ever happen.
Things you find in one area will influence what you do in others, so you will need to walls between roles in the business to be permeable. Customer service issues may need a marketing fix to set more realistic expectations with your customers. Conversion issues may be fixed by pulling data from customer service and passing it on to marketing.
It's good to start looking for problems and opportunities by specific, siloed areas. But you have to keep in mind that something you find in one area may lead into another. Everything is a part of the business, and these areas need to work together.
To get what you want, you have to know what you want. I'm not going to tell you to write out a 15 year plan, because that's dumb. I AM going to say that you should start out by knowing at least one big thing you want to achieve for the business in the next few months, so that you can lay out a viable plan to get there.
Objectives are the big thing you want to achieve.
KPIs are the main metric you'll use to determine whether or not you achieve your goal.
Targets are any specific numbers you want to hit to make sure you are progressing as fast or as much as you want
Step metrics are leading indicators that show whether you are on the right path. You won't always need step metrics, depending on how much is involved in the process to realize the objective.
Once you know what you want to accomplish, you know what data you need to be looking at on a regular basis to measure progress.
You can look at a LOT of food. But you can only eat so much at once. Pretty similar to your data. For each business objective or project you are working on, make sure you gather the right data for the project, and avoid fluff.
- Things that aren't related to what you're doing. Your email open rate matters, but it sure as hell doesn't come into play when you are optimizing a product page to improve add-to-cart rate.
- Vanity metrics. Having a site-wide low overall bounce rate is good... but it doesn't help you learn or improve anything. It doesn't really mean much.
- Attention drift. Keep just the information you need in front of you and do one thing at a time. You will be more effective with your time this way.
So for each project, figure out what actually matters, and look at it in a meaningful way on a regular basis. How can you get the data you need when, where, and in the format you need it?
Inside of Google Analytics, there is a "Custom Reports" feature that will let you create reports that show exactly what you need. That means you can take specific dimensions and metrics and create a stripped-down versions that give you exactly what you need. You can also create entirely new combos to give you a better look at things that might not be measured together in GAs native reports.
Dashboards give you access to the exact metrics you need in one place. They can be CEO/Owner style high-level overviews. Or they can be a drilldown into one specific area such as email marketing or device-based speed and performance. GA has a built-in dashboard tool, but there are better options.
I recommend Google Data Studio. It allows you to bring in any GA data you want, but also allows things like Google Ads, Facebook ads, and a range of other data importers for free. There are also free and premium services that will let you connect other data sources to GDS. That means you can have whatever data you want in a single dashboard, bringing together everything you need to see in one place.
Shopify isn't currently a free direct-connect offered by Google, but there are several ways to connect it. Most of the data you would need to pull from Shopify can come from GA with enhanced ecommerce enabled, anyway. GDS is a fairly new tool, and will likely keep expanding what it can do, and what free connectors it offers.
Automated email reporting
You can set up email reporting to automatically send out dashboards and built-in or custom reports to anyone on a one-time, daily, weekly, monthly, or quarterly schedule. This is great for keeping other stakeholders in the loop on a project without letting them dive deep into your analytics and coming up with too many (un)helpful questions and comments.
Asking good questions about your business and your website is the first step to making improvements. Since data answers questions, the questions have to come first if you want business value from analytics.
- Act as a starting point for further investigation or an optimization change.
- Have an answer that will mean something to you. You will either be able to act on or investigate the answer.
- Don't seek to just verify that your assumptions are right. You can start with an assumption, but look to confirm OR disprove it.
The first round of questions come straight from your business goals. What do you need to know?
- Where are the leaks in my store?
- What pages are making visitors leave? Where do they bounce/exit?
- Is the problem technical, design, or content?
- Which traffic sources are producing revenue?
- What gives the best ROI?
- Where should I put my money and double down?
- Which ones need work to improve?
- Which ones aren't worth it?
- Who are my ACTUAL buyers?
- Does my messaging appeal to them?
- How do men and women behave differently in my store?
- Are the people actually buying the ones that I think are buying?
- How many times/visits does it take for a visitor to buy?
- What traffic sources bring them to the site for the first and then follow-up visits?
- What pages do buyers see?
- How do they flow through the site?
- Which pages cause them to bounce or leave?
Asking next level questions
The first level of questions should lead you to ask more questions and pull out more information.
"So what?" will tell you whether or not you can take some sort of action. You can either investigate further, or you have a solution. If the "So what?" question doesn't lead to one of those, the information probably isn't very useful.
"Why?" should lead you on to understanding the cause of the event or trend you see in your data.
Useful: Firefox browser on Android phones makes up 20% of our traffic, but never converts. So what? We should look at user recordings for that segment and find the cause, because fixing that should boost conversions.
Not Useful: Our overall time on site is up 10% this month. So what? ....Nothing.
Firefox browser on Android phones makes up 20% of our traffic, but never converts. Why?
- Create a user segment in GA to see where these users leave the site. Maybe a specific page or part of the checkout process isn't performing well.
- Look at user recordings to see if something is going wrong. See for yourself what is happening with these users.
- Check page speed reports to see if you have long load times specific to this OS/browser combo.
Questions help you plan & prioritize your actions. If you have a leak in the checkout funnel, or a particular page is causing tons of site exits, that's something to work on ASAP. Following up with more questions helps you find the way forward to fix the issues you find.
Once you have the insight into what's going on, you turn that into actions and changes that make your site better. The entire point of this process is to decide what things to do that will actually drive your business forward and make more money. So far, we've collected data and learned from it.
But what do you DO with it?
Without data, you just know that you want to make more money. You don't know what actions will make that happen.
With data, you know things like:
- Product pages convert well, but that the flow between category pages to product pages is a huge drop-off because category pages have super-slow load speeds. You can go to a developer and tell them you need the category page load times lowered without sacrificing image quality.
- Repeat visitors convert at 3x that of first time visitors, and from your live-chat questions, you know they are confused about features of you products. So you need to improve your product descriptions to better convey information about your products. You could also experiment with marketing efforts like retargeting, email, and push alerts to bring people back the site early and often.
In other words, you no longer have to guess what you should do next. Using your data gives you a better understanding of what you should do to grow sales on your site. If you hire a developer or designer to work on your site, you have specific things for them to do. You can give them what you know from data as a starting point.
When you enact these changes, you already have analytics in place as a measurement tool. Instead of taking someones word for it that changing something will work, you can measure the effect. You can actually see whether KPIs like conversion rate, sales, and revenue go up as a result of the things you do.
Data-driven Shopify Growth
- Find and Fix Your Revenue Leaks
- Google Analytics Goals for Shopify Stores
- What is Conversion Rate Optimization (and why should you care?)
- What's involved in conversion rate optimization for a Shopify store?
- Heatmaps: Mistakes and Use Cases
- Finding Your Keywords When Google Analytics Says They're "Not Provided"