Does your website have an on-site search? You can harvest these queries in the Google Analytics Search Terms report and use that info to figure out, for example, what products your users are looking for (but cannot find because you do not stock it ). The Time Lag report helps you understand how long it takes people to build up trust with you and make a purchase. This info is the basis for decisions about how much to invest in remarketing etc. The users Flow report shows - at a glance - you where users land on your website and where they are most likely to drop off and leave.
over 1 year ago
No notes available for this episode.
Transcribed by Rugo Obi
The focus of today is on a special report that can turn your on-site search into on-site research.
Google Analytics has the ability to store the search queries happening on your website and then turn them into information about what kind of content your users want or what kind of products they're looking to buy.
I show you how to use that report and others in this video.
Lots of websites, including mine, have some sort of on-site search to help users find content they're looking for.
In my particular case, you're looking at it right here. There's this big search bar, and then a green button.
So I'm going to look for 'shipping' and see if this appears within any of the notes I sell.
And here we see a list of matches and there's a bunch of products that contain that.
Now, while it's certainly true that a fair few of my products contain the term 'shipping'. For example, this one 'Commercial Remedies' contains 'Lansat Shipping V Glencore' and 'Conflict of Laws' contains 'Armar Shipping V Caisse'.
What if I’m looking for 'Shipping law', that particular module? Let me see if I have that product.
This might be indicated by this search term that you're looking at here, and unfortunately there's no exact match.
Now, wouldn't it be great if you could harvest this searcher intent and figure out all the things that people are looking for on your website, but can't find? This is a very clear indicator of product lines that you probably should be stocking but haven't gotten around to yet, or didn't know were important commercially.
Well, it turns out that Google Analytics has this feature baked in.
You're looking at the
Search Terms report, which can be found under
Behavior > Site Search > Search Terms.
This report does not work out of the box, you have to do a little bit of configuration first.
So to do that, you go into the
ADMIN area, and then you select the
View Settings, and from here you scroll down until you see
Site search Tracking. You need to enable that. That should be on.
And then you need to set what
Query parameters are used in your website, in order to capture the search query. I'm going to show what this means by going over to the website again.
Here I'm back on my website again, and I'm going to carry out a search here with 'shipping' once more.
And then we arrive on this next page. I need to show you the URL bar in order for you to understand what's going on here.
You can see a query parameter has been added,
And what happens is, I've told Google Analytics that the parameter where the search term is contained is
I've also said there's another parameter query. That's because another part of my on-site search, well a different on site search, uses the parameter named
Now that we have everything configured, let's have a look at the actual data generated.
I'm going to select the
Ecommerce tab once again because I'd like to get some commercial data.
So, by default, everything's organized by the total number of unique searches. So here with the most frequently searched, which is 'company law', 198 times searched. And the value of each of those searches is 7 euros, this is a highly valuable search.
And then we get a bunch of other ones. In fact there's a total of, I think that's 28,000 total searches, and 16,000 different searches.
So maybe you want to view 500 at a time here, it's going to take a second to load. And you can see a list of all the terms. This is a very very long tail, as you can imagine.
How might you mine this data for actionable insights?
One thing to do is look for things that get searched for relatively frequently, but don't end up causing any financial transactions.
So we could do that now by going into
Advanced here, and then looking up something like search
Total Unique Searches, and we'll say
Greater than, I don't know, let’s make it
And then we'll also filter for number of
0, I'd say, and let's see what turns up.
Looking at the results.
For example, here I have
hume, and that was searched for 42 times, resulting in zero transactions, then media law, searched for it 35 times with zero transactions as well.
This indicates to me that either I don't stock these products, or I do stock them, but what's on offer is not very good, and therefore it isn't converting people and getting them to buy things.
In some businesses your customers won't buy anything from you on day one. They just haven't developed that level of trust, they need multiple interactions with your website before they are willing and ready to convert.
That's what the
Time Lag report is all about. You can find that under
Conversions > Multi-Channel Funnels > Time Lag.
The next thing you want to do is select the type of conversions you care about.
By default, Google shows all conversion types, but that includes both
Ecommerce > Transactions which have revenue amounts assigned to them, and regular
And these may have arbitrary financial amounts assigned to them, but it's not going to be super accurate compared to
And also, it'll inflate the number of conversions. I have 190
Ecommerce > Transactions here, whereas there's 10,000 of the
Time on Website > 3 minutes goals, which isn't that interesting really.
So I'm going to hit
Apply there and filter this data to just
Ecommerce > Transaction (It doesn’t seem to work. Let me do that again.) There we go. It's filtered.
Anyway, let's have a look at the actual data.
So if we look here,
Time Lag in Days, this is how long after someone visited the website that they actually converted.
In this case, 94 of those conversions -so roughly half of them- ended up converting on the same day.
The flip side of that information is that 50% of conversion values only occurs after the zeroth day.
And if you scroll down, you can see in fact that a lot of it happens, way, way, way later. For example, 12 to 30 days here, there’s 20%, and then there's another 10%, 60 to 90 days later.
So, you might be saying that's interesting, but what do I learn from this?
Well, due to the presence of things like remarketing campaigns and email marketing campaigns, the information that you might need to interact with the customer multiple times before they convert is actionable, in that you can spend more of your effort and budget, and so on, on these sorts of marketing activities, in order to have a customer visit you multiple times, then have that trust to actually make a purchase from you.
The next report I'm going to demo is the
Users Flow report. You can find that under
Audiences > Users Flow.
And basically what this does is show you where users end up landing on your website and where they go next, and most importantly, where they drop off, where they leave your website.
For example, the most popular landing page here or category of landing pages is this
law_cases stuff here. I'm getting a lot of traffic there, but 88.6%, you can see in the red there, are dropping off.
So, I should really analyze this and figure out "is there a way for me to keep that traffic on my website and bring them closer to monetization?"
I'm now on the corresponding web page on my website, this law cases- type page.
And if you browse it here, you can see there's not a lot of links leading people towards the commercial side of my website.
So I'm going to carry out an experiment and add some content here that will have some marketing potential.
I've done that now. I've added this yellow highlight box that advertises the associated product and has a little picture there.
So the test will be whether or not the addition of this box here will do something to reduce the amount of drop off from this particular category of pages.
Now, I don't know how that particular experiment will go, but my hope is that this number will drop, and I'd be able to notice that by comparing some recent period with a previous equivalent period.
That's all I've got for today. See you next week.