Recommend a web analytics tool

Getting started with web analysis: Selecting, setting up and using the right tool effectively

This article is your guide to finding the right tool for analyzing your website, setting it up correctly, and then getting the most out of it. Our author Britta Behrens from the analytics provider Piwik PRO will give you many useful tips from practice and, last but not least, explain what terms such as the bounce rate are all about or what a funnel analysis is good for.


Web analytics can be used to evaluate the influence of digital communication and marketing measures on the company's success and to derive recommendations for action. If the website serves as a sales channel, the performance and user-friendliness of the shop can be examined in more detail. In addition to classic business financial controlling, which is completely natural and necessary, you should therefore establish website controlling with the help of web analytics tools.

Many companies are not even aware of the valuable information that is available from it every day about their customers, which they could use successfully for their operational business on various levels (e.g. product development, marketing, sales, corporate communication, etc.). No expensive market research or test procedures are necessary, because the data is immediately available to whom it is stored centrally and evaluated in a targeted manner.

Remember: the interaction with the website is a direct communicative act on the part of the visitor with the company. This communication can be controlled and influenced. And the more you understand user behavior and intent, the better you can tailor your products and services to the individual user.

The website is therefore an integral part of a user-centered content strategy and provides valuable data for operational implementation.

With UPLOAD Plus you can read such articles as e-books or PDF. And that's just one of many exclusive advantages. Find out more now ...

1. Choosing a suitable analytics tool

Before you decide on a web analytics system, you should make a few preliminary considerations and prepare for use. Here are some suggestions:


Find detailed information on the websites of the individual providers. They often offer information material for download, present case studies and introduce reference customers. Gather a lot of information before contacting the provider. Research testimonials from independent third parties.

SaaS or self-hosted system

At the technical level, you have to decide who should look after the system. This applies to server hosting, maintenance and updates. Here it is important to find out whether this can or should be done by your own IT, whether a specialized agency is engaged or whether the responsibility of the analytics provider is transferred. This is a first step in the decision-making process as to whether it should be a SaaS tool (software as a service, i.e. a cloud service) or whether the system should be self-hosted by IT on its own servers. In addition to the technical component, legal and security-relevant aspects also play an important role here, which will be presented in the following.

Data Ownership and Access

The data that is collected from your website does not always automatically belong to you. Read each supplier's terms and conditions before making a decision. You should make sure that you have access to 100 percent of your data. Even if a contract is terminated, you should receive a final backup of your data. You should also make sure that no other companies are allowed to use your data for third-party purposes (e.g. advertisements, market research). Capture and use the amount of traffic you need for meaningful web analytics reports. It is recommended to use reports that are created without data sampling, as they only determine statistical averages. Make sure that you can use all the data collected and decide for yourself when samples of data are sufficient.

Integration with other tools

Check an analytics solution that can be integrated with other tools (e.g. Tag Manager, Optimizely, HubSpot, AdWords, SharePoint and many more). This is how you ensure smooth marketing automation. The better you can connect your marketing, collaboration and SEO tools with your web analytics software, the more efficiently you can carry out analyzes and campaigns. The workflow and reporting becomes simpler and more agile.

Advice, support and SLAs

Before signing a license agreement, take sufficient time to get advice. Test whether the provider has a suitable solution for your requirements and makes individual recommendations. Also clarify in advance in which form you will receive support and services during the contract period.


Use the opportunity to get an idea of ​​the software and find out whether the manufacturer offers a demo version. This will give you an impression of the tool's range of capabilities, get to know the user interface and user friendliness without having to use your own data.

trial version

Most providers offer a trial version for a limited period of time. So you have the chance to test the system with your data and to check the functionality. The success of a software stands and falls also with the user friendliness. When employees get on well with the user interface, they act confidently and more purposefully. People like to use the tool.

Software, IT and personnel costs

There are three components to consider for the cost of analytics software:

  • The license costs for the use of the software,
  • necessary IT costs (hardware / hosting / maintenance)
  • and the personnel costs for the employees who work with the software.

In the end, it is not the high-performance storage of thousands of data by the software that is decisive, but the subsequent interpretation. Here you should build up skills internally. The valuable analytics know-how can only flow into company processes and operational business if investments are made in competent personnel.

Get an overview of all costs incurred when using the web analytics system. You should only make a choice after you have determined all the expected costs. See web analytics software as an investment. They increase the company's performance and discover potential for optimization. Plan for the long term. Analytics software cannot be implemented and established in the company in just a few weeks. The use of the analytics tool should be aimed strategically over several years. It is a core element of your future corporate controlling. Take into account evaluation cycles during which you check whether the software continues to meet the requirements. A change should be carefully considered. Hence, it is better to be safe than sorry. So prepare the decision for the right system well.

2. Legal aspects, data protection

Web analytics systems are subject to data protection laws and guidelines. Before you implement the tracking code of your analytics provider and save data, you must conclude a so-called ADV (order data processing contract) with your provider. This regulates that he stores and processes the data on your behalf. The use of the web analysis tool is only permitted after this contract has been concluded. However, this is not the only legal measure that you need to take into account. The contract is only the first milestone.

Data anonymization

In Germany and the EU it is imperative that the data of the website visitors are anonymized. This is done by shortening the IP address before the data is saved by the web analysis tool. This function has to be integrated manually into the JavaScript tracking code. If this does not happen, the data has not been lawfully collected and may therefore not be used. If you have already collected data without an existing contract and without IP anonymization, it is advisable to delete the data immediately. The penalties for data protection violations will be increased significantly in the course of the new EU General Data Protection Regulation (EU GDPR), which will come into force in May 2018. Accordingly, you should have clarified all legal aspects before using a tool.

Cookies, opt-out and DoNotTrack function

In addition, the user must be able to prohibit website tracking on the website. This is the so-called opt-out function, which is also known from newsletter unsubscriptions, among other things. This should be clearly recognizable and easily accessible for the user. The ePrivacy Directive of the EU requires, among other things, that a cookie notice must be displayed when visiting the website, which the user automatically accepts when surfing the website or confirms this by clicking a button. This note is necessary because cookies are absolutely necessary for web analysis, to track the visitor's actions and, above all, to recognize him during a subsequent visit.

Privacy Shield

Finally, it should be mentioned that if you choose a tool from a US provider, you should make sure that it is certified for the Privacy Shield. It defines data protection standards between the EU and the USA, which companies in the USA must comply with when they save and process data from European citizens. You can neglect this point if you decide on a self-hosted solution or on a cloud offer that is hosted directly in Germany or Europe.

The Hamburg data protection authority has published a handout for companies on what data protection law must be complied with by companies in order to use a web analysis tool, in the special case Google Analytics. You can download the PDF from the authority's website.

An overview of the legal pillars that are relevant when using an analytics tool:

  • Federal Data Protection Act (BDSG)
  • EU General Data Protection Regulation (EU GDPR)
  • ePrivacy Policy
  • Privacy Shield (previously Safe Harbor)

Security aspects

In addition to data protection laws, there are also special security guidelines in many industries and companies. For example, in the financial sector, healthcare, telecommunications or government organizations, it is common for sensitive customer and communication data to be stored. Accordingly, the web analytics data should also be stored internally on the IT infrastructure's own servers. This avoids access by third parties and the data is only used for internal purposes.

When doing a risk analysis, you should weigh up the importance of owning and having full access to your data. Both SaaS and self-hosted analytics systems are common practice. Each company has to decide individually according to its requirements which system meets all the necessary requirements. At Internet Innovators and in Search Radar 62 you will find further articles from me on this topic.

Stay Up to Date: Handpicked Reading On Content Strategy, Online Marketing And More!

Register now and you will always receive on Mondays:

  • The latest UPLOAD content
  • Selected reading tips on other pages
  • Discount codes for exciting events
  • … and more!

No spam! There are already over 2,400 readers.

Further information on the content and data protection can be found on this page.

3. Preparation for the use of a web analytics system

With the decision in favor of the tool, the most exciting phase begins: The implementation and establishment of the analytics system in your own company processes. If possible, the preparations should take place parallel to the selection process for the tool. Because if you are already aware of the goals that are being pursued with the tool and that these have been formulated by the individual departments, the better you can make a selection in the decision-making process.

Set strategy and goals

The analytics tools on the market cover a wide variety of analysis methods. It is not the limitation of the individual tools that is a problem, but the almost unlimited possibilities of collecting and analyzing data. In the first step, get an overview of these analysis methods and create a catalog of requirements, which you specifically want to find out from the website usage data. You should consider beforehand which departments need access to which data and reports and which employees are responsible for the analyzes and reporting. Draw up an analysis strategy and set goals for the individual company areas and departments.

As on other occasions, you should define these goals using the well-known "S.M.A.R.T." Method. The abbreviation stands for: specific, measurable, attractive, realistic, timed. In this way, you set a realistic, measurable goal that can be achieved and must be met by a specified time. With the help of individually defined KPIs (key performance indicators) you determine the status quo and define a target value. In this way you measure whether you are falling short of, reaching or exceeding a target.

There is no standard formula for KPIs. Each company, each individual department defines this individually based on the data available, its own requirements and objectives. There is no run-of-the-mill plan that contains a specific list of KPIs and then has to be monitored and interpreted permanently. You need to find your own set of KPIs and set your own benchmarks. Align your goals with your current situation.

An example:

A shop has a natural conversion rate of currently 3%. This means that 3% of the organic traffic that comes to the shop via search engines or referrers is bought. A realistic goal would now be that traffic that reaches the shop via advertising campaigns (e.g. AdWords) must have at least a conversion rate of 3% and more, since additional expenses arise for the acquisition of these customers and customers are addressed through targeted targeting who have a purchase intention. If the conversion is below 3%, the campaign parameters (advertising medium, advertising material, targeting, keywords, etc.) should urgently be checked and optimized. Another sensible target in this case would be that the income of a shop, which was generated through ads, must at least cover the expenses of the campaign. The campaigns would be refinanced 1: 1. If you also take into account branding effects that the advertising campaign also achieves and has an indirect effect, the income can also be a little below the expenditure.

Reading tip: Here at UPLOAD Magazine you will find a detailed article on the subject of KPIs. In it: How to identify and select KPIs, create dashboards and much more.

Well-intentioned advice: Don't try to do everything at once. Establish an initial set-up of goals in the individual departments. When you have worked out a large number of questions and success measurements and can define and measure them on the basis of specific goals, prioritize them and start with the most important ones. Observe and analyze the goals and make optimizations based on the results.

Each goal is ultimately linked to one or more operational measures. After the data has been interpreted, there is a reaction phase and adaptation of communication or marketing measures. If you start with too many targets at the same time, you will literally overshoot. In the first moment you have a lot more knowledge, but you can never convert it into new processes or actions in parallel. This demotivates the teams and puts them under unnecessary pressure.

Let the success of the individual goals take effect and gradually implement the resulting measures. After you and the individual departments have gained experience, you can continuously expand the goals and analysis methods. Web analytics is not a project, but an accompanying company process. The individual goals and reports must be questioned and adjusted on a regular basis.

4. Analysis methods in web analysis

Web analytics tools offer a wide range of analytical methods and metrics. Not everything is equally relevant or different departments have a special focus. Admins and developers tend to observe the performance data, i.e. how quickly a website loads and whether the system is permanently available. The sales department controls how many leads can be acquired through the website and converted into customers. The social media team monitors the communication campaigns and measures how much traffic the external communication transfers to the website and how high quality it is. The SEO team evaluates its optimization measures and the traffic generated from them. It is therefore clear that a suitable set of metrics is selected for each goal, which reflects the development and success.

A selection is briefly presented here so that you can get a feel for the sense and purpose of these metrics and analysis methods.

Visits (session)

This metric indicates how many visits from users have taken place in a period of time. It is also called a session. A session comprises several page views and actions by the user and can be of different lengths.

Unique visitors

This number indicates how many unique visitors have visited the website in a period of time. The number of visits is usually higher than the number of unique visitors to a website, since visitors can start a session and visit the website several times.

Pageviews vs. Unique Pageviews

The number of page views indicates how often a page was viewed. For example, if a visitor visits the homepage three times during his session, it will also be recorded three times. There is also the metric of unique page views. Here the start page is only recorded once for the respective session.

Entry pages

This metric indicates which pages of my website are frequently accessed first. The home page is usually the main entrance gate to the website. Strong online communication, search engine optimization, advertising or good backlinks from other websites make subpages particularly relevant. Here it is often necessary to check which measures have led to the entry pages and, if these can be controlled, how high-quality the traffic from these sources is. If individual pages have high inbound traffic, which in the end leads to too few conversions or a high bounce rate, the causes should be investigated. For example, it could be that the entry page ranks for a search term in Google that does not fit the company at all thematically. It should be considered here to reduce this ranking so that no irrelevant traffic reaches the page, which dilutes the usage figures.

Length of stay

The length of stay indicates how much time a user spent in total on the website during a visit and on which pages they stayed for how long. In this way, you can recognize longer whereabouts and identify pages to which less attention is paid. When it comes to the dwell time, you have to consider that analysis tools only record the time measurement for a page when a subsequent page is called up. The time of the last page of a session is not recorded and not included in the total calculation. However, since this can often represent a relevant page at the end of the customer journey, one should try to record the length of stay. Manual intervention is necessary here with the help of event tracking. A time stamp can be activated by triggering an event on the page (see below).

Bounce rate

The bounce rate indicates how many visitors viewed a single page on the website and then immediately left it. The primary goal is of course to have the lowest possible bounce rate for all of your websites. After all, you want users to consume the website extensively. Depending on the type of page, a high bounce rate is negative and the cause should be investigated. The homepage should have a low bounce rate. With other page types such as a blog article, a high bounce rate is not per se problematic. For example, the user comes across a blog article through a search query that matches thematically. He receives all the important information he was looking for and ends the session. The bounce rate is high, but the user is happy.

Event tracking

Let's stay with the example above: In order to determine how many of the users really read the blog article to the end or ended the visit early, a so-called event can be integrated. It is determined when the event is triggered. With blog articles, it makes sense to measure scrolling whether the reader has scrolled 50%, 75% or 100% of the page. The implementation of event tracking triggers four actions for the user and measures the time. Events can be used to measure further actions that happen on the individual website without generating a new page view. This is how you record interactions such as clicks on buttons, forms or downloads. Another side effect, by triggering the actions, leaving the page is no longer counted as a bounce. The rate goes down.

E-commerce tracking

E-commerce tracking is an essential component for shops. Here you can recognize good and poorly performing products, recognize rising and falling sales at product category level, order quantities, average shopping carts and also receive a quota for abandoned shopping carts and much more.

Goal tracking

Goal tracking is an important tool for determining your KPIs and thus measuring micro and macro conversions on the website. Micro-conversions are, for example, downloads, completed forms for lead nurturing, newsletter subscriptions, visits to a specific landing page and much more.A macro conversion is a purchase made in the shop or a sales request. With the help of goal tracking, you can determine when a goal is achieved by the user and should be included in the reports.

Funnel analysis

An important part of the customer journey analysis is the evaluation of the "funnels" (in German "funnel"). They represent different steps on the website that a visitor should go through in order to complete a conversion at the end. They are also known as conversion paths. The funnels are based on a goal. A goal is defined once at the beginning and then you have the option of configuring the various steps to achieve the goal. You can use existing or new goals. Popular goals are: Visiting a specific URL, downloading a file, ordering a newsletter, sending forms or completing a purchase.

Now think about what the visitors to your website should do before the conversion or how you want to lead the visitors to the conversion. In addition to your own assumptions about how the visitors move in order to define the funnel, you can also draw on and use the click path analysis as a source of inspiration. In the reports, you should carefully observe how many visitors enter the funnel and actually complete it at the end. If the exit rate is significantly high, you should check the individual steps of the funnel.

Click path analysis

The click path analysis is an important part of optimizing your website and should definitely not be missing from your analysis software's repertoire. You analyze and optimize the conversion paths and contribute to improving usability and user experience. Similar to the funnels, the click paths are always based on a goal. Click paths visually represent all the paths your visitors take that led to a goal being achieved on your website. In this way, you can check completely impartially how your visitors are moving around the website and what paths they are taking until they convert to a goal. Here you can see whether your navigation and content concept for user guidance are well structured and understandable and whether the visitors follow the navigation elements you use (links, banners, buttons, forms, etc.) and take your intended path.

Campaign Tracking

Campaign tracking is one of the most important tools to record your online marketing and external communication measures and to evaluate the contribution to the success of your goals. You use it for performance and conversion analysis for the individual channels and content formats. With the help of tracking parameters that you attach to your website links, you get the perfect overview.


The referrer report shows backlink sources that are supplying the website with traffic. Here you can see the sources with the most traffic. Here it makes sense to analyze them carefully. Have the sources grown generically through good PR, communication and marketing? How do the users who come from these sources interact with my website and my offer? Is the traffic relevant to your sales or does it also contribute to brand awareness?

5. Dashboards and reports

Different dashboards and reports are of interest for each department. This means overviews of various values. With the help of dashboards, you can therefore quickly see your company or departmental goals and the performance of the website. Once created, it is not necessary to query the individual analysis methods again point by point. They are available immediately. With the help of widgets you can arrange everything clearly. As a tip, we recommend creating a dashboard with corresponding analyzes for each question. If you map too many questions in a dashboard, this can quickly lead to a distraction. Focus on the data that is of situational interest.

With the help of the reports you regularly document the development of your goals and can automate your analysis processes. Departments or management that are not actively working with the analytics tool can easily be supplied with information. With the help of the export function, the collected data can be processed for visualization tools and presented in an appealing way.

6. Segmentation

A very important tool in web analysis is the segmentation of the data. It is the most powerful tool to aggregate your analytics data in a meaningful way and to consider special target groups or channels separately. Without it, the data is completely unfiltered and has only limited informative value. Quantitative evaluations of the performance and the development of the website can still be determined: e.g. the total monthly traffic of a website, the number of all orders per month, the general conversion rate, the most visited entry pages, the new newsletter subscribers, etc. However, you get here only an overview of the totality of all visitors.

With the help of segmentation, you can qualify them and divide them into meaningful groups. This selection makes it possible to qualitatively evaluate different traffic sources, communication channels and campaigns. Classic segments are, for example, mobile vs. desktop users, Android vs. iOS users, geolocation for a country comparison, demographic characteristics, SEO traffic, campaign traffic, social media channels, etc. But that's only the tip of the iceberg: segmentation can be structured in extreme detail and can be used for numerous scenarios. In the analysis tools you have the option of combining all metrics with one another. Countless variations can be created with the help of and / or links.

Think about segmentation criteria that are of interest to your goals and make a list of segments. In this way, the analysis methods and metrics become significantly more meaningful and you can derive operational measures from the findings.

7. Sources of error and data discrepancies in web analysis

Of course, there are also sources of error and data discrepancies in web analysis. For example, if you use several tools in parallel in the evaluation phase to have a comparison of your data over the same period, you will find that the data differ from each other.

Not every tool measures exactly the same values ​​as another. This can have several causes. Some things are trivial, some require more in-depth know-how. It may be due to different data acquisition or a different definition of metrics. With a 5% deviation one is still within a statistically justifiable framework. If the differences are significantly greater, you should question the data and search for the cause. It doesn't always have to be an error. In a post on the Piwik PRO blog, a colleague of mine shed light on the subject.


After the web analysis is before the web analysis! After successful implementation of the web analysis system and the establishment of processes and workflows as well as the first milestones in the target assessment, it makes sense to initiate the next phase.

After you have gained sufficient routine and experience, you should expand and control your web analysis with the help of a "Tag Manager". A tag manager helps you automate processes and refine the analysis options. You will notice: If you dig deeper into web analysis and want to examine specific website elements or areas, you will have to regularly intervene manually in the website code. A tag manager is used to save valuable time during implementation and to minimize dependency on the IT department. This is controlled via a web interface and can be used independently by the responsible department that needs new tracking options. This becomes more and more important and relevant, the bigger the traffic, the more web pages the website contains and the more actions and elements have to be analyzed in order to determine and then evaluate campaigns and KPI data. In this way you can exploit optimization potential and become more agile in digital marketing. If you want to go deeper into the topic of tag management, I recommend my article "Tag Manager: How to take your web analysis to the next level".

You should prepare the decision for a web analytics solution well and allow sufficient time for the evaluation. You are making a long-term strategic decision. The better you define your catalog of requirements in advance, the easier the selection will be. Then the most exciting phase begins: establishing yourself in the company. I wish you success!

This article belongs to: UPLOAD Magazin 47

This issue is about monitoring and evaluating your key figures from website to social media. Which values ​​are important for what, how are they correctly named and how do you measure them correctly? Which tools can help you and how can you set them up individually for your purposes? And why don't all tools measure the same values? We clarify these and other questions in this issue

She is online marketing manager at Piwik PRO GmbH in Cologne. She is responsible for the German-speaking market and deals with web analytics, search engine optimization, content creation and distribution, SEM, SEA and social media on a daily basis.

Keywords success measurement, UPLOAD magazine 47