Determining Performance Goals

[00:00:00]
>> Todd Gardner: So, who gets to decide what is fast enough? And the short answer is, not you. Like, you don't get to say, four seconds is plenty fast for our website, we don't need to worry about it anymore. And you just deciding what's fast enough. Because your users might not agree with you.

[00:00:20]
There's three people, three concepts, that get to decide what's fast enough. And that's user experience, your competitors, and your SEO page rank. So if you look at user experience, to understand what user experience is telling you about web performance, you have to look at your business metrics, not your web performance metrics.

[00:00:46]
So what your business metrics are will depend on your unique situation, but it could be things like bounce rate, session time, how frequently things are added to cart. How frequently a cart is abandoned, what your conversion rate is, scroll length, you were talking about earlier. It could be any number of things.

[00:01:03]
It's how your business measures success of the website, and you have to look at those measurements. And then you have to relate those measurements back to your performance scores. So for example, here on the top, I have some web performance scores charted over time, my LCP and CLS and INP.

[00:01:22]
And on the bottom, I have some of my business metrics charted over time, bounce rate, and session time. And what I look for is, when does this correlate? Is there a time when something has gotten better or worse, and I can see a relationship with my business? In this case, when the CLS went up, my session time went down, in this example.

[00:01:47]
Or on the right, when my LCP went up, when it was slow to render the largest element, the bounce rate went up. And these both anecdotally feel like they make sense. If my page is bouncing around while I'm loading it, people are gonna get frustrated and they're gonna leave, and cause a lower session time.

[00:02:06]
And if it's feeling like it's taking a long time to load, which is LCP, the bounce rate would go up. But this is an example, I wanna underline the point that this isn't necessarily you, you have to look at your metrics and how your business metrics and performance metrics relate.

[00:02:26]
Because your users are gonna value different things and be affected at different thresholds. I wanna call back to this data that I referenced earlier from Cloudflare, that was tracking conversion rates and load times. And this is super interesting. This is how they found a correlation in their data between a business metric and a performance metric.

[00:02:48]
But correlation for them isn't necessarily correlation for you, and correlation is not always causation. This is my favorite graphic of all time, I find some way to work this into every single presentation I ever do. And it's the correlation between the global average temperature and the number of pirates in existence.

[00:03:09]
And it's true, it is correct data, and there is a direct correlation between the two. If we can stop global climate change by resorting to piracy, you would all think I'm ridiculous. And that's what this is trying to point out, is that correlation and causation are not the same thing.

[00:03:27]
We can learn from it, but don't take it as a golden rule that is 100% the cause.
>> Todd Gardner: Now let's talk about competitors, cuz your competitors are also a source for deciding how fast you need to be. If you are in staunch competition with other people, people compare your site versus somebody else, either directly or you're competing for the same search results.

[00:03:50]
You're competing for rank on the same terms, they are your competitors. And for a user to notice the difference between you and them, generally, you have to be 20% faster. And this is based on real logic called Weber's law, or the 20% rule. So here's some data. So, the graphic on the left, we have a number of dots in each square, and on the left there's 10, and on bottom there is exactly 10 more.

[00:04:19]
And glancing at it, you can tell that there's more on the bottom. You can look at it and be like, yep, there's definitely more on the bottom. But on the right, we're going from 110 to 120, which is still exactly 10 more. But you can't really tell. Like, visually, you look at that and you're like, I don't know if there's any more.

[00:04:38]
Because there's not enough, there's not proportionally, enough change. In general, it takes about 20% difference between two things for people to visually notice it, for people to feel it in performance, in any sort of way. So if we were to take that back to performance and our scores earlier, here was the scores I captured from Target and Walmart when I was building the slides.

[00:05:03]
They might be a little different as of what we recorded today, cuz they change day to day. But Target is only 4% faster than Walmart, and other than the fact that they're both a little slower, or they're both in the needs improvement category, which Google might care about.

[00:05:19]
As far as the user concerns, it's the same. User's not gonna notice 4% difference, nobody cares. Now, if I compare Target to Kmart, now, when I ran the same test on Kmart, Kmart has an LCP of 5.99 seconds, which is clearly slow. Target is 57% faster on LCP than kmart.com.

[00:05:43]
And a user will notice that, a user will feel that. A user will compare the two and be like, I don't want to stay here because it is too slow.
>> Todd Gardner: Now, the third thing that gets to decide is your SEO page rank. And if you're not ranking number one and you feel like you should and your content is just as good, your core web vitals are likely the cause.

[00:06:04]
And you need to hit these targets on these three metrics, according to Google, to avoid any penalties of your page rank. And so, those are the three forces that control how fast do you need to be. What do your users tell you, what do your competitors tell you, and what does SEO tell you?

[00:06:23]
Use those to inform you on how fast you need to be for your website.