Dealers will frequently ask for speed tests to be done on their sites, or do their own tests and ask for our feedback and help implementing changes that will improve their score. This article will help you run a speed test, respond to dealer concerns, and present your recommendations to the dealer for how to resolve some common issues.
Running A Speed Test
Even if the dealer provides us with a speed test they’ve already run, it is best practice to run one of our own. It is not uncommon for third parties and marketing services to run a test or present the dealer with falsified information to make their services seem more needed. Additionally, dealers often don’t know what parts of the test to attach/include to show us the whole picture or what items are specifically causing an issue.
GTMetrix is what we use to run speed tests. You will need to sign up for a free account to use it.
On the GTMetrix site, it will prompt you for the URL of the site. Paste the URL into the box and let it run.
Once the test has run, the results will be provided in multiple categories, which will be explained in more detail below.
Reading A Speed Test
Once the test has run, you will be provided with a lot of information, so let’s break down what is most important to look at.
Dealers will frequently look at their final “grade” and get upset, without understanding the context behind the score, or what can be done to improve it. However, once we’ve run this test, we can look at a variety of factors that the test tells us about.
Render-blocking resources
This tends to be the most “negative” category on tests of our sites. However, there are some elements on this list that we can not remove/change, and some items that speed tests will flag despite not actually preventing the site from loading.
This list contains the following:
Our global CSS file
The fonts used on the site
Bootstrap
Font Awesome
The dealer’s CSS file
JQuery
Modernizr
Drive It Now
Qualify Wizard
The dealer’s home scripts
Supplemental jQuery
You can see on the right hand side, each item has a “download time” that indicates how long each item took to download when loading the site. Items that are listed in milliseconds are typically not a concern – most users will not see an improvement between milliseconds. However, we can take a look at the items that are taking more than 1 second to load to see if we can do anything about them.
3 of these items are integral parts of the site we cannot remove (Bootstrap, CSS, jQuery). However, one of these items is QualifyWizard, which is a third party (which we will get into more in the third party script section). We should note this down while reading through the report to come back to when making our recommendations.
It’s also worth noting that although the rating for this section is “high”, the actual download time is not that high. Most customers will not notice or complain about a site taking 2-3 seconds to load. However, there are some small things you can check for the items that we cannot remove – for example, if you notice that their CSS has a lot of custom CSS in it that is commented out/no longer necessary, you could potentially take this opportunity to clean it up and make it smaller to reduce load time.
There are also some items that will be flagged as preventing render/loading that are not actually doing so. Audioeye is a common flag for this. Audioeye waits for the rest of the site to load to complete its own loading, so speed tests will often say that it takes 5 or more seconds for it to load. If a dealer complains about Audioeye being a loading issue, they may need an explanation about how it loads and why, although speed tests will flag it, it is not an issue that they need to worry about. (Additionally, although Audioeye is a third party script, we should not remove it as it is in place on all of our sites for legal reasons.)
Images
Speed tests often reveal that there are some large images on the site that are causing the site to load slower than normal. If a dealer is uploading their own images, they often do not think to size down or compress their images before putting them on the site. This can lead to some situations where images are flagged for being especially large or troublesome. In those scenarios, we should find the top handful of images that are being flagged and compress them (using TinyPNG or another similar service) for the dealer and re-upload during the process of the speed test. We should also avoid using PNGs at all unless the element needs to be transparent.
Once you’ve replaced the images with the new compressed versions, run the test again to see the improvement. Make sure you’ve noted which images you changed out in order to appropriately fill out your recommendations to the dealer (i.e., if the images they are uploading for their image rotator on the homepage are all too large, you’ll want to make sure you advise them to use a specific size/compress their banner images before uploading).
Third-party scripts/code
One of the most common causes for a noticeable slowdown on a dealer site is that they are utilizing too many third party scripts. Dealers often do not understand that implementing multiple products that are not intended to be a part of the site can cause issues. Because of this, it’s important to identify which third party scripts are causing the most slowdown, and encourage the dealer to weigh the slowdown of each one against the benefits they receive from using them. You may also find Google Tag Manager codes when searching through these lists – dealers may be using these containers to run additional scripts, which can slow the site down as well.
You can identify a third party script by the URL provided in the report. Most third party scripts have a pretty clear indicator as to what their origin is. For example:
The list above shows some third party scripts that were flagged as causing issues. We can examine the URLs to determine what their source is.
widget.kenect.com is Kenect (a text chat service).
googletagmanager.com is Google Tag Manager.
wsv3cdn.audioeye.com is Audioeye.
script.hotjar.com is Hotjar (a heatmapping service).
static.criteo.net is Criteo (an advertising service).
Some of the most common third party script concerns we point out to a dealer are:
Multiple GTM containers. In the example above, they have 7 different Google Tag Manager containers. We already have one set up for every dealer by default with most standard conversions, and while some dealers may prefer to have their own additional container they control, there is no good reason to have more than two. Seven is above and beyond what they need in any circumstance, so it would be best to remove most if not all (besides ours) of these.
Multiple text chat services. Many dealers utilize services like Kenect or RVChat to allow customers to directly text chat with an employee or chat bot to help answer their questions. This can be a useful tool, although it often slows down the site, but there is no reason to have more than one of them. In that scenario we should urge the dealer to pick one.
Multiple unidentifiable third party scripts. If we can’t easily discern a lot of the scripts they are using, chances are high that they are not reputable or good third parties for our dealers to use.
A generally long list of third party scripts. Even if we don’t know the full function of all of them, the more services a dealer uses like this the more their site will slow down. Having more than 2-3 additional services should not be necessary.
CLS issues
Coming soon
Making Recommendations
Once you’ve read through the report and understand the major issues, it’s time to let the dealer know what you recommend (or what you may have already fixed) in order to improve their score.
Some dealers are expecting more of a lengthy, comprehensive report – in which case, you can use this handy template to write up what your recommendations are.
Other dealers may just be looking for advice on a particular matter, or may have some idea of where the issue lies and want a confirmation. In that scenario, you don’t necessarily need to write up a full report – you can just let the AM/Support know what issues you saw/fixed in a more informal manner.
The most important thing is to ensure that you are honest but polite and informative with the dealer in regards to what changes you would make to see their speed improve.
Communication Examples
Below are some examples of how to (and how not to) communicate specific issues in your report. Generally, it’s important to make sure that you are polite, clear, and assume that the dealer does not know anything about what issues they have to make sure they have all the information they need.
❌ You have way too many GTM containers on your site. Get rid of them and the site speed will improve.
✅ Currently, you have seven GTM containers on your site. Typically, we would recommend only utilizing our single default container we have set up for you. Your site speed could improve dramatically if you were able to clear out or combined the function of some of these GTM containers. We recommend looking over which GTM containers you are actively using and allowing us to remove the ones that are no longer in use or are not necessary.
The second response clearly explains why seven containers is too many so the dealer can better understand our recommendation, and suggests some ways to find a compromise rather than demanding they remove all of them.
❌ There were a bunch of images on your site that were way too big and the wrong file format. Next time you upload them, you should compress them so they don’t slow down the site so much.
✅ We noticed that many of the images on your site, such as banner-front.jpg, were much larger than our recommended pixel size of 1600 wide. We took the liberty of resizing and compressing the images that were causing the most issues on your site, and saw your speed improve by quite a bit! We recommend making sure that all of your images are no larger than 1600px wide when you upload them, that you use JPG image files instead of PNG wherever possible, and that you run your images through an image compressor like TinyPNG before uploading to the site to make sure they create as minimal of an impact on your site speed as possible.
The first response does not give enough information about what was wrong with the images, which images were wrong, or how to fix it going forward. Our goal is to ensure the dealer has the ability to take care of this themselves, especially if they are uploading their own images, so it’s important to give them specific details about file size and how to compress images on their own. The second response also mentions that we helped them by changing some of their existing images to meet our standards – the dealer will likely appreciate that we went above and beyond to assist them this way instead of leaving it all to them.
❌ Why do you have two chat scripts? That doesn’t make any sense. You should get rid of one or the other, or preferably both, to fix your site.
✅ We noticed that you have multiple chat scripts – RVChat and Kenect. The more third party services a site utilizes, the slower it tends to be, and both of these services seem to fulfill similar functions (chatting with customers). We would recommend taking a look at which one is more useful to you/gets you more leads, and sticking with that one. This will also lead to less confusion for your customers as well, since they will not have to wonder which chat service to pick. The more sure the user is about what to pick, the more likely they will reach out quickly and get connected with a salesperson who can help encourage them to buy.
The first response is, of course, a bit too blunt – but it also doesn’t lay out for the dealer why they should reconsider their choices. It helps to encourage dealers to think about navigating the site from a user’s perspective to help them understand why something might not be recommended – in this case, as a user, seeing two different options for chatting in would be very confusing. Which one should they pick and why? Does it make the site look cluttered or unprofessional? Does one service have a very intrusive pop-up and the other one doesn’t? We don’t want to make a decision for the dealer, but we can provide our perspective and explain why one may be more off-putting for users than the other, or at the very least why having two can be confusing. They can use this knowledge, combined with how effective they feel the services are on their end, to determine which service they should stop using.
Assistance Expectations
Coming soon!