However, all of the email campaign’s links must be tagged so that all of the visits generated by your Newsletter can be correctly identified. When a link has been poorly tagged or not tagged at all there is the risk that visits will be allocated to the wrong traffic source.

With AT Internet’s EmailTagger application you can configure your email campaigns in just a few clicks. E-mail campaigns are tagged automatically by following these 4 quick and easy-to-follow steps:

1. Declare your campaign
2. Set the campaign’s send date
3. Download the email campaign source code
4. Retrieve the tagged email campaign

All that is left to be done is to automatically send the email campaign to the subscribers of your newsletter.

With the EmailTagger application, one of my customers reduced the time needed to configure and tag their weekly email campaigns. Instead of 2 hours per week as was the case previously, the customer now only spends 15 minutes on this task. Not forgetting any links and with no possible tagging error, efficiency has improved dramatically, meaning that the customer has more time to evaluate and monitor the performance of their email campaigns once they have been sent.

Florian T. - Web Analytics Consultant – AT Internet France

3 major constraints associated with analytics projects:

• Continual urgence which means a lack of time for integrating the solution, sometimes very short operational deadlines, and a need to access information immediately.

• Diversity of the different challenges:

- Varied and changing web technology and support.

- Companies have a wide range of jobs, goals, and needs for analytics solutions
(strategic/operational), directly impacting how the analytics solutions are used.

- Each online sector has its own characteristics (what do an online banking site,
an e-Commerce site and a media site have in common?).

• System reliability is a must in terms of the exhaustivity of data collection, ease of access to information and the coherence of the data displayed on the interfaces. The lack of reliability and the hypothetical security of some systems discredit the data which is available to strategic decision-makers and discourages operational decision-makers in their daily work. These security flaws may cause serious problems associated with private data.

At AT Internet we rely on agility both to overcome these obstacles and to help you carry out your web analytics project successfully.

Our approach is based on 3 pillars: Reactivity, Flexibility and Strength. Here’s why:

3 concrete responses

• Reactivity:

- First and foremost it allows you to implement your solutions in the quickest time possible, and with Soft tagging this is possible in the space of a few days. This technology, which considerably simplifies the technical phase of tagging, offers marketing teams more autonomy, allowing them to adapt or refine their analyses after tagging.

- It gives immediate access to information using real-time measurements (to the minute) irrespective of the volume of traffic analysed or the complexity of the request made.

- It guarantees reactive support in the event of a problem, as well as a team of expert consultants who are quick to react on all phases of the web analytics project: from support through to strategic consulting. AT Internet was recognised on an international level for the quality of its services in the latest Forrester study, placed ahead of all the other Web Analytics solutions in terms of customer satisfaction.

• Flexibility:

- Measuring all channels (web, apps, mobile sites, videos, social media etc.) to analyse their uses and their evolutions.

- Adapting to the objectives of each company by allowing companies to create their own specific business variables and to accurately target their analysis via a 100% configurable requester and real time segmentation engine, available for all variables without sampling.

- Placing online data at the heart of a company’s web marketing ecosystem and CRM, by providing automatic connections and authorising two-way data flows by API.

• Strength:

- Guaranteeing system performance in collecting, processing and accessing data, as well as a large capacity for scalability depending on the volumes of data to be processed.

- Providing a solution which conforms to the legislation in force, certified by key monitoring organisations, with a highly secure technical infrastructure.

Summary:

AT Internet’s company strategy is based on these 3 pillars. Our latest developments, such as the Soft tagging interface, our Full Rest API and our new requester are the first innovations from a long series which be will revealed over the forthcoming months.

With the custom sources feature, it is possible to create source categories (organic and marketing) that are best adapted to your needs. Group together social networks, or all other types of sites that you would like to analyse so that traffic coming from these sources can be isolated from the rest of your traffic sources.

Custom sources are configured on your interface via Tools > Sources and campaigns > Organic sources (or Marketing campaigns) > Custom.

An e-Commerce customer was trying to identify the traffic generated by their ‘Price comparison’ sources. He placed these sites into a custom source to measure the audience coming from this organic traffic source.

Olivier D. Web Analytics Consultant – AT Internet France

The creation of the QR Code custom sources has allowed one of my customers to bridge their print campaigns to their online returns. This gives both the customer’s marketing team and myself a means to comparing conversions from paper version campaigns with conversions from online campaigns.

Ben M – Web Analytics Consultant – AT Internet UK

A customer wanted to isolate traffic coming from different news sites following a different number of articles about him. We suggested that the customer create a custom source which would allow him to isolate the excess traffic, coming from this series of articles, from the rest of the organic traffic coming from the site.

Rémy B. – Web Analytics Consultant – AT Internet France

This post contains an overview of the event and the EU Cookie Legislation discussion (particularly as it relates to the UK’s implementation of the directive). The second post will discuss how to get more management buy-in from your senior management team and Board of Directors, providing concrete examples of ways to get their attention. The third and final post will review multi-channel attribution and the philosophy behind attribution models.

Before we dive in to the cookie discussion, we think it would be good to have a quick review of the Digital Cream format for those who aren’t familiar with the event.

Digital Cream is an invitation-only event for Digital Managers of some of the world’s largest brands. Attendees are free to share their struggles and successes with each other without having to worry about anyone quoting them due to the event being run under Chatham House Rules which encourages open and frank discussions. Here’s a sample of just some of the companies that actually participated on the web analytics table :

• Walt Disney
• Turner Broadcasting
• DK Publishing
• Eurostar
• Net-a-porter
• Dyson
• Canon Europe
• Marks & Spencer
• Uniqlo
• Glaxo Smith Klein

Discussions are held at large, round tables with one moderator, one sponsor representative and 8-10 attendees. The moderator guides the discussion, making sure everyone has an opportunity to speak and the sponsors are there to add their expertise to the topics raised by the attendees.

Now, on to the first major topic – cookie legislation.

EU Cookie Legislation in the United Kingdom

The impending deadline for the implementation of the UK cookie legislation was a hot topic on all the tables and created some of the liveliest discussions. The infamous ‘Table 13’ was purely dedicated to the topic of cookies, but we also came to the same basic conclusions in the discussions at our table.

First of all, it goes without saying that according to the current UK government guidelines you should have opt-in cookies implemented on your site(s) no later than May 26, 2012. [1]

However, if that’s not possible due to development schedules, budgets or you simply haven’t started the process yet, you should consider the following recommendations.

1. Perform a complete cookie audit immediately. It is essential that you understand exactly what’s happening on your site and why before you begin to create a cookie management and permissions plan.

a. Remove cookies no longer being used. Many people mentioned that when they did audits they found loads of old cookies still being placed for software and services no longer being used by the company. This is a privacy risk and is also creating unnecessary weight on the pages. Get rid.

b. Consolidate as many cookies as possible to reduce the overall number of cookies being placed on users’ computers. The fewer cookies the better.

c. Move to first party cookies where possible. This should not have a major impact on your web analytics, no matter which tool you use. Since the intent of the legislation is aimed at third party cookies, the fewer of them you have on your site, presumably the lower the risk of exposure.

2. Update your privacy policy.

a. Use ‘plain English’ (or simple, plain language in your native language) to explain your privacy policy – particularly as it relates to the use of cookies. Clearly explain why it’s necessary for you to use cookies and what you do with the data. (Personally, I like to use the ‘grandmother test’ i.e. would my grandmother understand it? If not, keep working on it until it is simple enough for non-technical people to understand.)

b. Describe the cookies you are using. It’s not necessary to list each and every cookie but general statements like, “We use cookies to measure the effectiveness of our website so we can make future improvements for you.”

3. Develop your company’s plan.

a.    Create written documentation (PDD, PRD) of the project outlining how you are going to make changes to your site to obtain opt-in permission from users for the use of cookies. The plan should include all the normal things you would expect to see in a project plan:
- Who – team member
- What – list of requirements, new site design, wireframes, etc…
- When – the development/deployment schedule
- How much – budget figures
- Etc…

b. The Marketing department should develop the initial design, verbiage and overall plan before sending it to the legal and technical teams for approval. This establishes ‘ownership’ early in the process and allows better control of messaging and delivery.

In summary… no one really knows exactly what to do. It is up to you and your company’s management team to decide exactly which course of action is best for your organisation to take. The prevailing opinion is that you should be doing something instead of just waiting to see what happens.

Have you implemented a project similar to this one and would like to share your story with us? Have you identified any other leads or recommendations? Feel free to share your experience with our community.

As I said earlier, nobody knows exactly what to do. A discussion on this subject may just enlighten some people.

The next instalment will discuss ways to get buy-in from senior management for web analytics – from seeing the value of the analytics reporting to getting budget for additional staff to support the work.

What does that mean for you?

As Jean Marie Camiade told us in a previous post, the principle holds in the use of the secure https:// protocol and the encryption of navigation data. Google’s main services, Gmail, Google+ etc… include an impressive number of users, and are constantly growing (1 billion Unique Browsers in May 2011 on all Google services [1]) so the impact of these changes will be more and more pronounced over time.

Of course it causes consequences for many SEOs and web analysts because request performed by that “logging” user on a Google service would be in a secure protocol and therefore keywords could not be collected and used directly for optimization purposes of content, positioning in SERPs (search engine results page) of Google, or even customization in real time depending on the keyword typed.

On March 5, 2012, Michael Safyan, a software engineer at Google, said on the official Google blog “Inside Google Search” that these changes will now expand to other local Google domains around the globe including google.com, google.co.uk, google.fr, etc…  In the same article, companies are also being encouraged to move to using https:// protocol gateways.

For security reasons when someone navigates from one secure website (https://) to a non-secure website (http://) no information can be captured and therefore there’s no possibility of collecting, say, the referrer.

Maybe because of Google’s advice to switch all sites to https:// protocol, some companies have mistakenly interpreted this to mean that if they switched to https, that they would then be able to capture the information.

This is not the case.

First, Google uses https. Second, Google removes keywords typed by users by using a script that websites are not be able to receive in SSL.

“Google uses a proprietary ‘hack’ to remove the keywords from the referrer. Otherwise you wouldn’t even get the “not provided” notice. Google removes the keywords using a script so that even SSL sites can’t see them.”[2]

So, there’s no need to rush to implement this change on your website. You can evaluate impact on your own data if you analyze the keywords on Google on March 6^th. Traffic keywords from unknown ([-]) should be increasing from that day.

Note: A study dedicated to Google research in https will coming soon in the “Barometer of search engines” and will complete the previous.

New positioning: Agile Business Decisions

Our main task is to provide our customers with the right information and to help them make the correct business decisions, in other words help them reduce their expenses, optimise ROI and increase profits.
Our plan is to continue this task today, with the same reliability, except that agility will be instilled at the heart of our customers’ decision-making processes.

Agility as a company strategy

The close relationship that we have with our customers has made us aware to what extent our customers’ working environment is continually changing and how it forces them to constantly adapt to such changes and to be agile so that they can make the right decisions.

Agility is an already tried and tested concept at AT Internet: from product design and development, to our commercial strategies, and the organisation of our product offer as well as our multi-business and multi-channel expertise.

• Agility has always been at the heart of our innovation process. We work hand in hand with our customers so that we can keep up to date on their ever evolving needs. In return, our customers help us guide our developments and roadmaps.

• More recently, agility has been a strong component of our commercial approach. We consider our customers’ new projects as a gamble on our shared progress. By offering to accompany and assist our customers on the long term, we will build on their improved skills and competence, whilst at the same time subjecting our own development to new and powerful challenges.

The agile solution

Our range of products and services has been developed so that it can be adapted to all evolutions. Our products and services can be used as a basis for any feature or training module which in the end will form THE solution that is both relevant and suited to our customers’ projects, maturity and business sector. This made-to-measure solution may change over time at the same rate at which our customers’ skills and competences also evolve.

Today, we would like to take the concept of agility and make a real company strategy from it to help our customers meet their goals and objectives.

A new website

The work involved in reorganising our product range and establishing our position has led us to developing a new graphic charter and revamping our different communication supports.

Origiami was chosen as the graphic concept for the new site. This Asian art of folding well illustrates how agility and knowledge can transform a simple piece of paper into the most diverse and sophisticated of shapes.

The Agile Business Decisions campaign will be portrayed by the colibri (hummingbird). In our opinion, this fast and very reactive bird is a good ambassador for agility, to the benefit of your business decisions.
We wanted our new website to be clearer and at the same time more concise. We hope that it will be the correct platform for displaying your messages. The web agency “WSB”, who we have been working with for several months, has paid particular attention to simplifying user browsing.

This blog post is also an opportunity for us to thank the agency for their involvement in our project.

Why not log-on to the website www.atinternet.com ?
Please feel free to leave your opinions, comments and suggestions on the blog.

For companies new jobs need to be created, new skills integrated, and there is also a new channel of communication to get used to; for better or for worse.

But more particularly, the social web is also about new realities:

• Controlled communication is but a distant memory

The power of speech is at the hands of Internet users. You need to make sure that Internet users have a space in which they can leave comments and ideas, and where they feel listened to and free to express themselves.  Create spaces for open discussion according to your goals and objectives.

• No presence strategy, no success

Giving the power of speech to Internet users is good, but being present, identifying oneself and communicating (somewhat differently) is even better! Your presence strategy must be defined according to your expectations and your target market, and must also respond to well defined objectives. Thanks to these first criteria, you will be able to define which platform is most suited to your community management actions.

• In the space of a few hours your online reputation can be a success…or not

Be careful. Listen, act, react, be attentive and more specifically, reactive. Monitoring tools are essential in becoming aware (often in real-time) of the conversations currently taking place. They provide information on the latest trends, the general feeling towards your brand, channels and the most active authors. When configured and used correctly, monitoring tools can help detect risks and anticipate future actions.

• Measuring ROI remains fragile

In this domain your only salvation will be the use of KPIs. As is the case in web analytics, the use of KPIs is essential. The different KPIs may refer to the engagement, and interactions around your brand, the range and influence of authors/sites which talk about you, conversions or leads that your actions have been able to generate, or the general feeling associated with your brand.

To resume, take the time necessary to create a realistic approach: an approach associated with a solid strategy, efficient monitoring and of course the required human investments. This approach will make your introduction to the social web successful and will allow you to take advantage of it.

What the social web must not change is your approach and your work as a web analyst. As is the case in web analytics, the approach to analytics must be strict, developed for the long-term with suitable tools and a solid strategy.

Do you require any assistance or suitable tools? If so, we can help. Please feel free to contact us.

• Firstly, a computer prediction machine capable of detecting trends and alert points from a sophisticated algorithm, and artificial intelligence which will automate and simplify the detection of relevant information.

• Secondly, a web analyst, capable of understanding the context in which a piece of information can be found, capable of prioritising the data that is available so that effective recommendations can be made from it.

We at AT Internet, editors of technological solutions, stimulated by numerous creative engineers, should be the ardent defenders of our leading solution. We should try and get increasingly closer to finding the magic solution which will raise questions, provide the answers and apply the solution. However, it’s nothing of the sort. We are not resigned, rather the opposite, as the progress of our decision-making solutions is vast and there are many subjects which still need to be dealt with (for example, the smooth inter-connection of the different traffic sources, the visual representation of decision-making data, the convergence of video feeds or even social and mobile supports). We are therefore continuing to search for the magic solution and develop at a more sustained rhythm to improve the solution’s autonomy.

However, we remain convinced that the real hero of tomorrow is not (and should not be) a computer programme, but man himself. This belief is based on several different reasons:

1.       The instability of the perimeters studied. In a nutshell, websites change on a daily basis (for example: changes in ergonomics, the addition or closure of certain sections, new formats), meaning that there is no sufficient data history to rely on to automate predictions. This problem is then aggravated by the long-nature of the tagging in place, and by the reduction of the volumes studied, brought about by the essential segmentation of traffic audiences.

2.       The complexity of the analysis context. The different elements which can directly or indirectly impact a site’s audience and the interpretation of Internet user behaviour are extremely difficult to work out and prioritise. For example:

a.      The weather,
b.      The latest news,
c.      The relevance of a traffic acquisition campaign,
d.      A new unexpected link on a partner site, or a new indexed page,
e.      The general technical performance of a site or of a particular page,
f.       Competition,
g.      Market trends
h.      etc.

We are working to integrate and cross all of these sources in our reports. This can be done quite easily for the weather, and we have been doing it for some time now with the technical performance data generated via our Observer module. However, let’s be realistic, it is somewhat utopian to link all of these dimensions together.

Computers have beaten man at chess, but the environment which the web analyst works in is infinitely less stable than a chessboard.

Why is total automation not desirable?

1.       Where should intelligence be placed, and the comprehension of different phenomena observed on the site? Suppose that (even if we don’t believe it) machines are able to detect alerts and correct them automatically (via custom content, a change in the value of a keyword for example), is it desirable for a site manager not to be aware of the problem and not be able to understand the solution? What would happen tomorrow if, for x reasons, the technology in place is no longer sufficient and must be replaced or changed? Do we want self-learning to be kept only for analytics tools or behavioural targeting? [1]

2.       If we extend our thoughts a little, we can see that there are ethical limits as far as automating analysis algorithms is concerned (which indirectly affects recommendations to be made). In an amazing article published in the New York Times, Eli Pariser, perfectly illustrates the dangers associated with the obsession of relevance and customisation, supported publicly by certain giants of the Web, using the same principles as our democracies, and the intrinsic richness of the Internet. This short article deserves to be read in its entirety, but its conclusion is clear: “It is in our collective interest to ensure that the Internet lives up to its potential as a revolutionary connective medium. This won’t happen if we’re all sealed off in our own personalized online worlds”.

A consultant’s ability to analyse and to place things in context, capable of understanding the complex environment of a company (human, technical, commercial, marketing) remains by far the most precious tool available to take advantage of any analytics strategy. Of course, without any high quality data (reliable, readable, quickly accessible and easy to interpret) the analyst will be kept in the dark, but this data, delivered as raw data or processed in a programme (the most intelligent programme in other words) will not be useful, and will even be dangerously powerful.

We are continually trying to find the balance between the technical solution and human assistance (via our consultants or our partners). We are passionate about developing faster, more intelligent programmes to contextualise our data, and from this data improve the decisions made. There is no mistake about the decisive importance of a consulting service to develop reports, define relevant KPIs, interpret data and suggest effective actions.

Web analysts still have a few good days in front of them, before leaving the way clear for Skynet Terminators, which, should I just add, have a notable lack of finesse.

Online ad expenditure will increase its share of the ad market from 15.9% in 2011 to 21.2% in 2014. The need for media channel attribution modelling is growing tremendously.

What you will get out of this post:

• How to identify consistent and cost-effective channels to produce revenue or brand awareness
• How to benefit from understanding the interactions of various channels
• How to determine the average time to purchase to look for improvement opportunities

1- Which channels produce results and which ones only produce traffic?

According to a survey conducted by Forrester on the Purchase path of online buyers in March 2011, nearly 50% of online transactions followed multiple exposures to web marketing efforts. Today marketers can follow a customer’s journey as they open branded email, click banners or paid search terms then continue on to order products or services. When done with proper scrutiny, this type of analytics provides marketers the opportunity to think beyond the traditional “last-click” measurement they typically attribute to marketing programs.
Of course, the first step is having a measurement tool that attributes sales to specific channels. For effective optimisation, you must be able to see the contribution that your different channels have on sales.

Once you can measure sales attribution, it’s time to begin optimising. A key point to agree upon is that a channel’s value is greatly affected by the quality of the marketing placed into it.

You want to make sure you are putting forward the best piece of marketing possible to give each channel its best opportunity to provide results. Work with your agency and solution provider partners on collecting benchmarking, keyword research and any other data relevant to your business and target audience segments.

2- Using Common Customer Paths to Optimise Channel Interaction

According to Forrester , “Fifty-seven percent of the interactive marketing professionals surveyed measure results for each channel, but less than half, only 28%, measure the influence of one channel on another”.

The problem with this “swim lane” approach to measuring channel effectiveness, is that consumers don’t interact with a brand or make buying decisions in straight swim lane style interactions. It’s also possible that a customer interacts with several channels before making a purchase.

Customers seldom follow a “swim lane” when interacting with channels en-route to making a purchase.

Due to the fact that customers can often interact with several touch points before making their purchase, the traditional “last channel” approach to measuring contribution to conversion falls short.

Focusing only on attributing sales to the last channel can mask some of the hidden gems that truly drive sales. By creating a more advanced attribution analysis model, you may find that something, which is less expensive like non-branded search terms, actually accounts for most sequences that lead to a sale. Finding out that a lower cost channel is an important sales driver can be a great boost to your ROI.

For example, AT Internet’s Channel Optimizer allows users to see up to 15 channels in the sequences leading up to a sale and even gives users a detailed access report per order:

An example of distribution of conversions, sales turnover and the acquisition of new customers on different sequences of channels.

Identifying key channels thanks to detailed information on converted sequences.

When you set out to optimise your spend on online channels, a good place to start is with the ones that contribute the most.

This may seem like a cumbersome task, but the cost savings that can be gained from recognising over attribution to expensive publishers, or over use of expensive marketing channels such as pay-per-click is worth the few extra man hours spent on doing the research.

Online intelligence, just like business intelligence, is an investment in streamlining your resources to achieve optimal results.

3 – Reduce the Delay between Channel Interaction and Purchase

Before the global adoption of exchanging money for goods and services online, marketers were constantly engaged in a battle to speed up a customer’s purchase cycle. Typcially the five stages are 1. Awareness 2. Information Search 3. Alternative Evaluation 4. Purchase Decision 5. Post-Purchase Behaviour.

The typical struggle was to speed up the cycle because if there was too much latency, especially between stages 2-4, it was increasingly likely that the sale was lost. Even in today’s 4G high-speed way with my iPhone society, marketers indeed continue to aspire to reduce the time between their advertisement and its intended purpose of creating a sale.

Luckily, with the proliferation of web analytics, data is now available to see the duration between a customer’s first interaction with your ad and the completion of an order. By using this information it’s possible to take action on channels that aren’t moving your customer through the purchase cycle.

As an example, combining data from AT Internet’s Sales Tracker and ChannelOptimizer Tools will give you the insight necessary to begin building an analysis model to help understand your customers’ purchase cycle.

By pairing up order and access details it’s possible to see the channels accessed, the duration between those access points, whether or not the user is a returning customer and other order details. Once paired up and placed into a spreadsheet you can begin to understand the average latency periods of different types of customers.

Following through with this advanced use of the data, you might see something like the fact that average durations between channel interaction and purchase go down after an influential blogger writes a post about your industry or product. With this information you can schedule your ad buys more effectively when you know that your customers are more likely to be closer to the purchase stage of their purchase cycle.

Partnering with solution provider’s like AT Internet you are able to request custom reports to be built to really get the most out of your data. Also with our growing list of online marketing partnerships AT Internet allows you to have increased convergence of your online marketing data.

Written by Michael Nemirovski and Carole Pero

The doctrine itself is often very elaborate, well controlled by the elite, showing only a partial, generalist version enhanced with every positive quality possible, whilst limitations and drawbacks are well hidden.

The stated objectives are mostly clear and understandable, and respond to the real needs shared by many. “The marketing packaging” is designed to convince people that the doctrine is the only way (or at least the best way) to achieve such objectives.
It is this process which has imposed the pre-eminence of data, and of figures in particular, on the analysis of such data.

“Data driven” in question

The idea of placing figures at the heart of decision-making is attractive because it is based on the premise that “the numbers do not lie.” The choice of these figures and their interpretation, however, are two questionable key points.
If the wealth of online data makes the concept credible, it masks the absolute necessity of a rigorous analysis in exploiting such data.

I briefly addressed this issue in the series “The job of the web-analyst” and “Context is Critical” but this article is complementary because it deals more with the heart of the problem.

From “data-oriented” (where the company is active) to “data-driven” (where the company is passive), these are two terms which are often confused.

The “Data Driven Enterprise” in question

No-one would deny that when it comes to implementing a corporate strategy a company needs to set targets, measurement methods for collecting reliable data, and accurate dashboards containing relevant KPIs (Key Performance Indicators).

We should remember the most important thing in guiding a company is not the data but rather the analysis carried out from them, just as a doctor should not treat the symptoms of a disease but rather the disease that causes them.

Let us be objective: who can give me the name of a company which cannot be described as “data driven”? Who knows a company which cannot rely on data, on figures to develop its strategy, draft goals and monitor performance using dashboards?

We just have to keep in mind how goods delivered by the caravans of yesteryear (before writing was invented) across the desert were managed. The courier was given a “pot” of clay, which was sealed by the sender, containing a stone per parcel to be delivered. The recipient would break the pot to check the integrity of the delivery. Already “data driven”?
Two different interpretations can be considered: either a system of data structuration (with metadata, semantics, etc.), or all-powerful figures…
…and that’s exactly where the problem lies: it is most often used in the latter of the two.

The dictatorship of numbers

There are two problems here:
- A skilled analyst can (voluntarily) make figures say almost anything and everything.
- An unskilled analyst can (unintentionally) deduct almost anything and everything from these figures.

As far as interpreting numbers is concerned, two concepts are crucial: the quality of the analytical process itself and the objectivity, impartiality and ethics of the analyst.
By simply reading the profiles for the position of Web Analyst, you will see that these two notions are almost never taken into account.

On the other hand, the benefits of services such as AT Insight or hand-picked certified partners provide clients with experienced consultants who analyse data on a strictly unbiased and impartial basis, ensuring the objectivity. One of the consequences is that sometimes it is uncomfortable to announce some conclusions, but it is our duty to do so.

The figures, which are the inspiration behind important decisions, must be accompanied by a contextual analysis (method of acquisition, calculation, scope, context, interdependencies, etc.). Nothing is more risky than making decisions on mere numbers, should they be sophisticated meta KPI without the contextual analysis.
Similarly, a dashboard is not a decision-maker, it is only a support tool for making-decisions (as discussed in the White Paper “Dashboards, a GPS for the Manager”).

I will conclude this article by expressing the wish that business analysts finally be seen for their true value, so that the analysis will take precedence over the figures.

This optical illusion highlights the dangers of interpretation: as you may have guessed the lines are straight, therefore parallel, although the context (the position of the black and white squares) leads us to believe the contrary.