Web analytics for business
You invest in marketing money. And you certainly want to know how many clients come in and how much is one given client worth. Better SEO or Facebook?
And can place a banner on the site X for 1000 UAH per month? Or it is better to turn off banner ads, it's all the same confusing little ...
An article on web analytics will not answer these questions. It will give you an understanding of how to find the answer yourself and receive this answer regularly.
The material is intended not for analysts but for entrepreneurs who need to understand: How to optimize the budget for advertising? How to check your marketer. There will be no big grooves and instructions on how to set up, all this can be found by Google.
Why web analytics business?
The possibilities of web analytics for business are not limited to the following. And to start you have enough
Increase website conversion
Conversion is the ratio of visitors to visitors who bought. If you present the site as a clothing store, then the number of purchases affects:
- The number of people who pass nearby;
- The number of people who enter the store;
- The number of people who come to the racks with clothes;
- The number of people who try on;
- The number of people who go to the cashier and pay.
Each step is a microconversion, which leads to the ultimate goal - the purchase.
People visit the site in the same ways. It is important to find them and measure the conversion rate - how many people out of 100 who tried on will go to the cashier.
Conversion rate increases by finding anomalies. For example, people who come to the site from screens with a resolution of 1920 pixels in width buy twice as often as the average. This means that the site for them looks different. You need to understand how to duplicate it for other sizes.
Case from practice:
We made a landing page for a pharmaceutical product. While optimizing the conversion of the site, they noticed that on screens up to 1180 pixels wide, the girl in the background image of the first screen does not fit into the screen. The presence of a girl increased the number of people who scrolled further 2 times. After such an observation, the background image was altered, an A / B test was performed, which confirmed the hypothesis.
The second example: making cuts, we noticed that the conversion of girls from 8 to 11 pm increases by 50%. What to do with it? The site will not do anything. And advertising in this period of time, girls can show more.
Please pay attention to the conversion rate of the site. Increasing it from 0.75% to 1.5% can be much cheaper than doubling advertising budgets.
In addition, as someone said: “An indicator that is regularly monitored is beginning to magically improve.”
Identify more profitable traffic channels
For example, you invest $ 1000 per month in each of these traffic channels:
- Facebook Ads
- Google search advertising
- Banner advertising
One of them brings customers for 10 thousand dollars a month and the other two for 15 thousand.
Obviously, 15 is more profitable than 10. If the tenfold return on investment in marketing is OK for you, then 10 is also beneficial. Analytics allows you to measure these numbers and compare the channels with each other. To do this, the easiest way to make a sign with a list of traffic channels, the cost of each and the number of customers that they brought. Ideally add revenue. This can be done by properly setting up Google Analytics and integration. And it is possible in the usual table monthly set of data by hand. And that works.
It is important that you determine for yourself the allowable cost of attracting a client. And if you have several products, then for each product.
And they understood how much a client costs for each traffic channel.
Keep the cost of attracting a client within the limits of acceptable!
Hypothesis Testing
Hypothesis is the assumption that by making X you get Y.
For example: "And let's raise the order button above, because no one is crushed there."
Analytics allows us to test hypotheses. An experiment can either confirm or disprove the hypothesis.
Case of life:
We tested two hypotheses at the same time:
- Increase the size of input fields and submit buttons.
- Reducing the number of fields.
It would seem that both items should give an increase in conversion. Just in case, we made 4 options:
- Original
- Full set, enlarged fields
- Abbreviated Dialing, Original Field Size
- Abbreviated set, enlarged fields.
It turned out that the concern of the manager and the company was important for the audience of this site. And they perceive the existence of the right questions as a concern.
Increased fields - definitely increased the conversion by 30%. But the decrease in their number - reduced conversion by 20%. Therefore, after the experiment left the original set of enlarged fields.
Advertising optimization
Having an analytical system is critical to performance marketing (or engineering approach to marketing). For example, a specialist in setting up contextual advertising can optimize campaigns only knowing the data about purchases, preferably with a cost. Similarly, remarketing, targeting and other promotional tools. Properly set marketing task is necessary and can be digitized, measured and controlled.
Adjacent areas
Data source for marketing
Analysis of the audience of the site and its segments allows you to identify additional data that will help you competently plan the entire marketing of the company. From the portrait of the target audience, its geography, content preferences to the correction of the product line.
Data source for advertising channels
Analytics systems allow us to gather audiences for, for example, remarketing on Google Ads and Facebook. From the simplest case “All visitors to the site” to combinations of the type “Who looked at the contact details of the manager, did not order anything from us and have been on the site more than four times”.
What is web analytics
Summing up, we can say that web analytics is a system of tools and practices for optimizing a website and advertising channels in order to get more profit.
The goal can sometimes be different, but these cases either boil down to profits or are very rare.
What does a web analytics system consist of?
Metrics and KPI system
Analytics allows you to measure and what to measure it - you decide. Of course there are indicators that are tracked in the default reports. They are undoubtedly important. At the same time, each business has its own tasks and it is from these tasks that you need to build a system of metrics.
What exactly to track?
I advise you to build a chain of actions of the buyer from contact with advertising to money on hand. With all the branches and options.
In real life, everything is much more complicated, but for understanding the general picture this is enough. Most likely the lion's share of buyers will go one or two ways. We need to understand now what indicators need to be monitored. What most affects the amount of money on hand? Average check? The number of checks? Conversion from viewing a product card to an order? Number of site visitors?
Select the key performance indicators of marketing. They are monitored almost every day. In turn, KPIs directly depend on the values of all indicators in the system. These indicators can be tracked for example once a month.
Customize analytics
Now we need to configure the analytics system.
UTM tags
For the correctness of tracking the source of the transition to the site I advise you to use YuTM tags. This is a special type designation.
? utm_source = content & utm_medium = it4u.ua & utm_campaign = lunch-analytics & utm_content = aticle which allows the analytical system to correctly determine where the visitor came from. Naturally, Google Analytics independently knows that the person came from Facebook. The problem is that in Facebook you can have ten kinds of activities. From paid advertising, maintaining your own group to working with opinion leaders and guest posts. That is why each type of activity is marked with its own “tag”.
Developments
In order for Google Analytics to understand exactly when the target action takes place in some cases, it needs help. Things like viewing the page or staying 60 seconds on the site can be tracked without special events. And some actions need to be flagged. This is done by calling the function gtag ('event', 'event_name') ;. The analyst will see the event that the function has initiated and mark it in himself.
Goals
The site generates many events. Not all of them are equally important. Those events that we track are set up as goals. A goal is an action that is important for us to monitor. This way we can see the number of actions performed in the reports.
Application Forms
Events need to mark all application forms. It is desirable to start filling out the form and the fact of sending. Pay attention - not the fact of pressing the button, but the fact of sending data. This is often done by inexperienced developers. As a result, the number of orders in analytics and CRM may not match.
CRM Integration
For completeness, it is also desirable to integrate the analytical system with the CRM system. So that when buying a CRM, a new event is reported to the analyst, which will continue the sales funnel
Import costs
Also now it is possible to import costs from Facebook and other advertising channels directly into Google Analytics. You will be able to see not only the conversion but also the entire picture by cost.
Result analysis
Advertising optimization
Correctly setting analytics advertising specialists will operate with the necessary data to optimize it. Much better to focus not on clicks but on revenue, right?
Associated conversion - what is it?
I'll tell you by example. The person found you through contextual advertising Google. I looked, read and forgot. Then he was caught up with remarketing on Facebook - he again got on the site and remembered you. I read your blog, subscribed to the newsletter. A month later, one of the letters hooked your client to live and he went to the site, ordered your service. By default, email reports will appear in all reports. What about remarketing and context? You can make the erroneous conclusion that they do not work. In order to make the right conclusions, there are attribution models in which the contribution of each channel is distributed fairly. How exactly - you can choose.
Seasonality
It is important to consider weekly and annual seasonality for correct analysis.
Source: https://habr.com/ru/post/440878/