Average time per visit handle with care

The average time per visit is one of the analysts favorites metrics.
The objective of this metric is reporting how sticky the website for its visitors is. The way most people use it is taking it directly from the average time per visit report. One of the most important things for a good analyst is first of all having an in depth understanding of the analyzed data. In online, all the information require a technical understanding besides the statistical understanding.
So, what does the average time per visit metric measures?. In most platforms average time per visit is calculated dividing the total time the website was navegated over the total amount of visits. What kind of insights would you be able to get from that information? In my opinion…none. why? Because you are measuring heterogeneous things together. All the navigated time including the people that entered your site, found out that it has not the required information leave the site inmediatly. If the quantity of non qualified users is relatively high will influence your metric at the point that the average could tends to cero.
A lot of people that has a website that requires a minimun navigation time get surprised when the average time per visitor is ten percent the required time for watching the welcome video/message. But at the end is completely logic, that metric is not usefull for that, it can not tells you the average time per visit of your site users when you are also including “lost users”.
So the main recommendation is as a first step using the non-bounced traffic segment, this way you can separate the “lost users” from your target users getting a much usefull metric.
But, what happen in cases where because of the analyzed system (website) a high bounce rate is not only not bad but even something good? Then you can switch thrr non-bounced traffic segment by the repeated visit segment.
This should not complicate your analysis unless you use external benchmarks, something that we never recommend. Compare yourself with your yesterday’s yourself, it would me more reallystic and challenging

It’s not you…it’s me

Computers are great in repeating while humans are very fool on doing that. On the other hand, humans are awesome creating while Computers are completely handicap on that field.

Normally we tend to think that some day will appear an analytics solution that will do anything for us, even making decisions. We try to think and behave as computers instead of taking advantage of our Human Nature. We try to assign unanimated things the human status…like in the following and hilarious video…

The thing is, computers can make decisions on new events. When something that never happened before appears computers are not able to deliver an efficient insight. Your Web Analytics Solution will only able to provide you information, the same information that your mental model will analyze and generates an insight. But don´t worry, it’s great! Human potential is amazing.

If you expect from your Analytics tool more than what it can gives you, you will be very disappointed…it’s like when you girlfriend told you…It’s not you…it’s me. By the way, take a look at this interesting metric…

Flowcards information on the go

Since I´m working in Analytics I´m trying to get an analytics model that allows me to understand and analyze each particular system in a properly way…as something alive.

The current Analytics models are great to analyze pictures of the system, or “movies” but about each part of the system, with an holistic and not systemic vision of Analytics.

The holistic vision says that if you optimize each part of your system you have an optimized system, while the systemic vision says that the key is not on every part of the system but on their interactions. You can have a beautiful landing page but it is important is if it (just to give an example) drive people to buy. With that in mind I was working on an analytics model that allows people to:

1- Understand each part of their system (organic traffic, paid traffic, website (the subsystem website), etc).

2- Identify the bottle neck (the resource preventing the system to get more results).

3- Find out the best way to action on that bottle neck.

4- Measure results. Get the information to feed the information system.

5- Identify again the new bottle neck.

The result was a model that I called Flowcard. This is the first time Continue reading →

Want a benchmark? You better get a mirror

We, people, love benchmarks. We love that someone tell us if we are doing well or not.
A benchmark is nothing but just a point of reference that give our metrics the necessary relevance that make them useful. There are two types of benchmarks and I wanted to highlight this because even when they look similar they are completely different. So, let’s go for it…
1- External Benchmark: This is the most commonly called “Benchmark” and the one that I don’t like…at all. Actually they don’t want to be face to face with me . Each website, or digital project as a whole has it own particular system and in that system each part interacts with the rest in a very particular way. So how can we compare two different systems with the same metric? Two eCommerce websites could have a completely different average pageviews per visit even when both performs well. Two Social Networks could have the same average time per visit even when one of them is very successful and the other one sucks.

2- Internal Benchmark: If you Continue reading →

Call it as you want just help me make decisions

It’s somehow funny but for some reasons we (people) use to generate multiple systems from something that has just one objective by generating some fake objectives. In this case it’s regarding the activities related with generating and providing information for decision making. This kind of activities are provided today by different disciplines (that in some companies do share departments and in some others don’t) like Market Research, Web Analytics, Business Intelligence, CRM (part of it), Data Mining, and so on. So is getting harder to understand what’s the limit of each discipline. So the people from those disciplines fight for hours arguing which information must be generated by the Market Research department, by the Web Analytics department, by the CRM deparment, by the BI deparment, and so forth.

So what came into my mind is why should I (as a CEO or business owner) have several departments that share the same damn objective? Nobody cares how big your ego is and Continue reading →

Analyze scenarios, not metrics

If you want to really understand what it is happening in your project I recommend you to analyze de scenario and not the metrics per se. An interesting way to understand it is like when you are about to buy a new alarm for your home. The expert ask you all about your house, your family, if you have a dog or other pets, etc. After that tells you how many sensors (Measuring tool implementation) will your house (online project) need. After that each sensor will be measuring all the house movements and if a deviation appears you will be informed (alert).

The above mentioned example shows how important is the “system” to make the metric make sense.

Having 80% of bounce rate doesn’t mean a thing unless you understand how does your website work. If you do have a blog with the entire note at the home and most of your visitors visits every note you post, then you should expect a high bounce rate, actually a high bounce rate could be a good metric for you.

So think on metrics as something that can’t be analyzed in an isolated way because lonely metrics are sterile…

Reporting Levels – Less is best

Companies are systems composed by different areas and levels. Those areas and levels perform a very important role in the organization that will (or not) allows it to achieve (or not) its objectives. Said that it is important to understand how vital is the information system flow.

Even when some people may think that this post is mainly focused on big corporations, it is not. Information flows will determine how effective is the decision making process into any company no matter how bit it is. If you have a 10 persons company and the 10 are making decisions with a 30 pages report, then are all of them losing very valuable time in understanding a love of information that was not generated with that person (or area) in mind, then will not fit their needs.

So, first step, please, pleaSE, PLEASE, it is extremely important that you understand your company structure, which means, levels and areas that compose your organization as well as the interactions among those parts. What is the role of each area?, how it help the entire structure to reach the system’s goal. Every single area and every single person in a particular area is important for the company (otherwise why are you paying that extra salary?) so the better they work the better the company results will be.

The following Analytics 2.0 Chart called Backbone Chart explains in a very graphical way how it works.

The Backbone chart shows Continue reading →

Information is not an insight itself

I know, you probably read the header and said, I know it is not, however sometimes the obvious it is not that obvious so let’s go for it.

Information is just set of data ordered and processed in an specific way that make it relevant for decision making. However the human brain is responsible for the magic of understanding a situation using this information. A tool will never be able to do this.

Humans performs very poorly with repetitive processes but are great creating things based on information understanding.

To break the wall and invade the “Insights” arena it is necessary to ask yourself what the information means in terms of the System you are analyzing. So, having a bounce rate of 80% means nothing, the fact that your conversion rate is x% doesn’t mean a thing by itself. The information must be placed in the correct part of the system to understands it meaning. This way an 80% of bounce rate in a landing page with a video that is not being measured is not a bad rate, actually you must expect something like that.

An Insight is not the information itself, is the magic produced when the Analyst place the information in the correct place of the system and the answer appear in front of his eyes. Did you ever feel that?

Towards Qualified Analytics Professionals

I always enjoyed my analyst role. The idea of trying to find answers in a mess of data and information sounds incredible challenging to me, it is probably (I guess) like an adventure game or something like that (I’m not a expert gamer so I don’t really know the name of those games ).

I’ve been an Analyst for more than 14 years and I’ve never got bored, actually the thing is getting even better. However for some reason some people believe that if you want to growth as a professional you have to climb up into the organization, which means, changing your role, which is in my opinion absurd…let me give you my reasons.

Whether if you are an Analyst, a Coder, or any other profile, you require some particular skills, mental model, and even values that are probably completely different that the ones you need to perform another role. Which means that you can be an awesome Analyst that enjoy his job with passion and when you get the “so important” promotion as a manager becoming a weak (and sad) Manager.

So, in my opinion the normal way of climbing up as (for instance) an Analyst should be becoming everyday a more experienced and qualified professional (you can call it Jr, Semi Sr, Sr, Mega Super Sr. and so on). I have a lot of reasons why:
1- You MUST enjoy your job.
2- A brilliant analyst must be paid for his performance. So if the person generates value to the company, the company must pay for it, you don’t need to call the person “Manager” or “VP” for that.
3- A brilliant analyst is not necessarily a brilliant Manager and the brilliant Manager is as necessary as the analyst.
4- All the companies need brilliant and passionate people on each position.
5- Since the company is a system (a group of things interacting together with a common objective) every single person in the organization is responsible for accomplishing the objectives, ergo, every position is crucial.

Think about it…

Integration, where are you?

I’ve been my last two days at Adtech San Francisco and I was very surprised (and not) because none of the Web Analytics Tools where there. Google was there just with Adwords, Yahoo! had no booth, Omniture had no booth neither, Coremetrics the same. Ok you can say, they are big so they don’t need to be there? So, should I think that Adwords is small? You can also say that some of the tools are free and they are part of other business (like Adwords and Google Analytics) and here is when we return to the begining. Information is what makes you generate forecasting scenarios. Without information you are not able to understand what is gonna happen in the future (with and acceptable level of confidence), if you are planning without information from the past then you are just guessing and I’m not sure that guessing could bring you to a acceptable level of confidence.
On the other hand companies like Omniture or Coremetrics where no present there neither, and they charge a lot of money for their services. This has really no explanation to me. I know that were some other analytics events on the region those days, but come on, we (Analytics professionals) are always talking about integration and cannot even share the same space with adveretisers and publishers? Please, is not about Analytics, or Paid Search, Or Organic search, is about Business and businesses require all those things togheter.
So why none of the biggest Web Analytics Companies had no presence? No idea…
Finally there are a lot of new Web Analytics Tools companies all over the world, it was a great opportunity to fill the Web Analytics emptiness left by the other companies… but nothing happened…I’m really confused.
We (Web Analytics professionals) are always saying that clients and agencies are not taking Web Analytics very seriously, but how would they take it seriously if we don’t do it?

Analytics 2.0

Much more than just tagging

Category Archives: Analytics 2.0