I came across a tweet the other day that read: "Device traffic split for last week: Desktop 41%, mobile 35%, tablet 24%" ...and it points to one of my pet-peeves about modern analytics.
We are still basing our analytics on single device specific hits. This is not how the world works. First of all, we now live in a multi-device world, so dividing up our traffic isn't the right to do. It will mislead you.
But that's only half the problem. Think of it like this: If I tell you that I just sent you an email, what device will you choose to use to check it on?
See the problem?
You don't actually make that choice. You don't choose to check your email on a specific device. You just check it on whatever device that happens to be the most convenient at the moment.
It's the same in our analytic tools. For instance, I checked Google Analytics and it is telling me this:
- Total unique visitors: 372,036
- Unique visitors from desktop: 218,462
- Unique visitors from mobile: 117,802
- Unique visitors from tablet: 36,217
So, how many are using more than one device? Well, according to GA, none... none?!
I'm sorry, but this is just wrong. Google itself has found that 90% use multiple devices sequentially, and 67% shop only across multiple devices. So how can GA be telling me that *none* of my traffic is using multiple devices?
Now, I fully understand the technical problems here and why this is so, but that doesn't change the fact we are getting false reports.
"But, wait-a-minute", you say. "Isn't it useful to know what percentage of my traffic that is coming from each device type. For instance, in knowing what to focus on?"
No! No no no no no no!!!
You can't do this for three reasons.
Reason 1: You are not selling your products to 'visits'; you are selling your products topeople. If a person is using three devices he is counted as three visits, but you still only getone sale.
Reason 2: In term of future strategy, your traffic comes *after* whatever you implement.
I wrote about this back in 2012 in "The Mobile Shift is Already Here". You can't wait for your traffic to 'go mobile' before you start to have a multi-device strategy. That's not how it works.
I see this so often with brands. They look at their analytics and learn that maybe only 4% of their traffic is coming from mobiles and tablets. They think "Oh, mobile is not that important yet. Let's wait until we get more mobile traffic before we change our site or make our website mobile friendly".
If this is how you look at it, you are never going to get any mobile traffic. Your strategy is not about what your audience is doing today. It's about what you want them to do tomorrow.
And if you want to have more mobile/tablet traffic, then you need to embrace mobile and tablet design first.
One example of this is from this site. Here is the difference before and after I redesigned this site to be responsive. You will notice that before I redesigned it, I practically had zero mobile traffic, and that is despite that my audience is part of the first-movers when it comes to multi-device behavior.
But look what happened when I launched the new responsive design that works across all devices. Almost immediate my traffic patterns started to change.
So looking at your device split measurements has little meaning for what will happen tomorrow.
Reason 3: As I already said, people do not choose to use a specific device. They just use whatever device that happened to be most convenient one at the time. So it makes absolutely no difference to your overall strategy if you have 15% mobile traffic or 55%. Your job is to be 100% relevant, 100% of the time.
Being mobile is not a decision. It's the absence of a decision. It's the freedom of not having to decide what to use before you do something, but just do it whenever you feel like it.
But, isn't it relevant to see the device split to track how things are changing? Well... yes-ish.
In theory, it's very useful to see that after creating a mobile optimized site that your traffic starts to reflect that, like what I did above. But there is a problem with this.
We still only measure single devices, but people are behaving in a multi-device way. In other words, it's a highly misleading measurement.
Consider this: This is what devices my analytics is telling me that people use today. I have about 370,000 unique visitors in total, divided across desktop, smartphones and tablets. I might think this is important because if I can just boost my tablet based activity, I can increase my overall traffic, right?
Well, no. Because this is wrong. The trend that we are seeing is not that people are dividing themselves up into groups. The real trend is that we are becoming more and more 'multi' every day.
So what's actually happening is this:
Instead of having 370,000 unique visitors across three device types, I probably have closer to 200,000 unique visitors, with a substantial part of them using multiple devices.
So what do you think happens if you boost your mobile and tablet traffic? Do you get moretraffic overall, or do you just get a more multi-device traffic?
Knowing the ratio of multi device usage is critical. Without it, it's practically useless to look at device measurements.
Sure, from a trend perspective, it's very interesting to watch how we are using more and more devices, and the interactions that take place between them. But as a publisher and a brand, it's not really a choice.
Take this site. Only 9% of my traffic is tablet based, so I might conclude that this traffic wasn't that important and only focus on desktop and mobile. But that would be a big mistake. Because if I give people a bad experience on tablets, it reflects on all the other channels, as well.
You can't choose not to support one device. They all have to be good because you don't control when, where or how people might need you.
It's like having hot running water in both your kitchen and your bathroom. Sure, you probably use more hot water in your bathroom, but that doesn't mean that it's any less important in your kitchen.
via:http://www.baekdal.com/insights/stop-thinking-about-devices/
No comments:
Post a Comment