Episode 38 - Why You Should Ignore the Shopify Speed Score

The Podcast

Show Links

• Pingdom

Help the Podcast

Transcript

Hey.  Scott Austin here. 

In this episode, I'm going cover page performance for your Shopify store. 

The net of this episode is that you should ignore the Shopify Speed Score and not worry too much about page speed in general.  Here's my reasoning.

Let me get started by diving into the Online Store Speed functionality that Shopify recently released into the Shopify Admin.  It can be found under Sales Channels > Online Store.  Simply put, I highly recommend that store owners ignore this metric.  There are a number of reasons why I recommend ignoring it including:

• The Shopify Speed Score is based on Google's Lighthouse Performance Score.  Now you would think that a score coming from Google would be something to pay attention to.  But in this case its not.  And that's because the Lighthouse Score is very technical and makes many theoretically impactful recommendations.  But in reality, most of the recommendations are out of your control.  These recommendations could be followed by Shopify or your theme developer or your app developers.  But there's very, very little that store owners can do to remedy the technical recommendations.
• The concept of a score is deceiving.  When I think of something that gets a score in this way, it makes me think that a perfect score is possible.  When I see a score, say 84, I think that's a B.  And an A or A+ would be better.  So let's look at the Lighthouse score of the top 10 websites by traffic.  This list include Wikipedia, Instagram and Yahoo!.  The top 10 sites have an average score of 59.  Google search gets an 84 and YouTube, which is owned by Google gets a 45.  Only Wikipedia gets above an 85. So by putting this score on a 100 point scale, it deceives us into thinking that most web sites are failing.  When the reality is that getting an A is just impossible.
• Shopify's relevant comparison in the Shopify admin does not appear to be accurate.  I just looked at 60 of my clients scores in the Shopify admin.  The lowest was a 9 and the highest was a 58.  Ironically, the low and high are two separate stores owned by the same company.  The average is a 28 with most stores clustered between 14 and 32.  Now all of these are real, live, revenue-generating stores.  Shopify provides a relative indicator along with the store that is one these three statements:

1. Faster than similar stores
2. Same as similar stores
3. Slower than similar stores

• Speed score is not an indicator of ecommerce success.  There's a lot buzz on the internet about how every fraction of a second of performance improvement yields significant improvements in revenue.  But that doesn't appear to be true.  Let's first look at the top 10 Shopify stores by traffic, because revenue numbers are not published.  Well these stores have an average Lighthouse speed of 18.  Let me give you the numbers of each of them in order of score from highest to lowest.

1. USA Tuan - 67
2. Cettire - 23
3. B33 - 18
4. Kith - 18
5. JP Hi-Fi - 12
6. Pura Vida - 12
7. ColourPop - 10
8. AIRugaib Furniture - 8
9. Fashion Nova - 6
10. Gym Shark - 5

1. Amazon - 61
2. Etsy - 61
3. eBay - 45
4. Lowe's - 33
5. Target - 28
6. Wayfair - 26
7. Macy's - 23
8. Walmart - 18
9. Home Depot - 10
10. Best Buy - 8

So hopefully, I've convinced you that the Shopify Speed Score is not a useful metric.  Let me shift gears now and give you some useful ways that you can measure site performance.

The first metric to look at is your store's conversion rate.  Now, conversion rate doesn't directly measure page speed or size.  But it does measure your customer's response to your site.  If you are adding functionality to your store that negatively impacts the speed score, but your store's conversion rate continues to rise, then speed score be damned, you are doing the right things. 

Here's a way to think about it.  Let's say you sell cars.  Now on your car website, you add one of those visualizers that allows people to choose colors and options and see what the car will look like.  Now that configuration app is going to load more slowly than a generic product page.  But it provides a much richer shopping experience and will improve the conversion rate.  In this case, building a slower page actually improves revenue.

And its also important to note, that page speed is more important on your landing pages than it is on the customer's 3rd or 4th pageview in the session.  A fast landing page let's the customer know that've landed on a live, functioning site.  If you do have pages in your shopping experience that are slower like the car configurator, don't use them as landing pages.  And also consider adding a loading graphic, which helps the customer understand that something is happening and they'll just need to wait a moment.

Now this is not to say that page load time will never become an issue.  It can.  So here's an alternative way to measure page speed.  I look a single metric - load time.  Note, this is not a score, but the measurement of how many seconds it takes to load a page.  The tool I use for this is Pingdom, which I'll link to in the show notes.  Unfortunately, Pingdom also provides a grade.  I'm going to encourage you to ignore that.  Don't think about page speed as something you continuously tweak.  Think about it more as a pass/fail.  If you pass, great, move on.  There are other things that you can work on in your business that will provide a better ROI.  So what is passing for load time.  I say anything that is 3 seconds or less is passing.

Now, if your page load time is over 3 seconds, Pingdom offers two other metrics that are useful in troubleshooting the issue.  One is the total page size of all the HTML, Javascript, CSS, images, etc used to make the page.  If this is over 5 MB, then the size of these assets could be the issue.  And the biggest contributor to page size is images.  So your images may need to be optimized.  Another metric that Pingdom shows you is the number of requests.  This is a count of all of the files used to build the page.  And believe it or not, that can easily be over 200 files for a single page. 

So let's say you are trouble shooting your page that takes 5 seconds to load.  If the page size is under 5MB and there's less than 200 requests, then the most likely contributor to the issue is a poor performing app.  One common issue is that a single file can take a long time to load.  This happens when that file is not stored on a Shopify server and that other server is just slow to respond.  Here's how you can check for that.  In Chrome, open the DevTools using the F12 key or Control-Shift-I.  Then go to the Network tab at the top.  Then refresh the web page.  You'll see each asset for the page show in a list with a bunch of data about them.  Watch the Waterfall column.  If you see a row with a relatively longer bar than the other rows, that file may be the culprit.  Determine what is calling that file, is it an asset you added.  Or from an app?

So let me summarize my recommendations on how to approach page speed for your Shopify store:

1. Ignore Shopify's Speed Score
2. Track your conversion
3. Check your landing page's speed with Pingdom
4. If your pages load in less than 3 seconds, move on to other business priorities
5. If your pages load in more than 3 seconds, dig deeper.

That's it on this topic.  Thanks for listening.