1. Introduction
Which role does Web Analytics play for companies these days? Evangelists and tool lobbyists kept repeating that Web Analytics systems can provide insights into web site performance, and that it would replace gut feeling with facts as a more solid base for decision making. That still is entirely true, but over time the perspective has shifted and enhanced significantly:
The traditional angle for discussions about Web Analytics is a technological one: What can be measured? Which tool and/or measurement method is superior and accurate? What is the level of measurement we should apply for our web site? And: which are the key performance indicators that our company should focus on?
These questions are all indicating a stable perspective on things, based on solid technology, on business objectives, on the presumption that Web Analytics could contribute to replace organizational uncertainty with knowledge.
But: As the markets are changing we need to re-open the discussion about Web Analytics as source for insights. Two big topics are seen here: How the business is run and how the users are seen.
Web Analytics tools are measuring web sites, thus the questions need to be formulated with respect to and with perspective on web site contents.
So the first big question can be: How can we determine how well our web site content is accommodating our business goals?
The second look goes to the big question: What can we learn about our users when we look at Web Analytics data? Or then, with perspective to the measured entities: How well does our service or content offering accommodate for our users’ needs?
There are plenty of ways to answer these questions. Here’s mine.
My first hypothesis is:
1. Companies are institutions created with the purpose of neglecting change!
You may not have followed this, but at the moment there are plenty of traditional companies in Germany going down the drain: “Quelle” (a mail order company for consumer goods, 82 years in business), “Märklin” (building model railroads and other toys, 150 years in business), “Rosenthal” (porcelain, 130 years in business). Countless more.
In the three mentioned cases the infamous “management mistakes” were spotted as the root cause for these companies going bankrupt after having been around for long times.
It would be too short-sighted to argument that any of these companies would still be around if only they would have incorporated the computer in its various aspects: the Internet as a sales channel (Quelle), as a playground for computer-aided model railroad landscape design and the sale of accordingly customized packages (Märklin), or as a means for lean production, using CNC-driven robots to paint the porcelain (Rosenthal). Here’s why:
Every economic endeavour is forced and driven by its environment to incorporate certain things they cannot afford to leave out. You can call this “a permanent reality check”.
Every economic endeavour, on the other hand, throughout its operations, starts to build up a static counter force, something like a “culture” (or, although the word’s a bit worn out: an “identity”) which is considered to reflect the very core of the value proposition of a company in a market.
At the same time companies tend to atomize their own operations within the market and to look at changes within their business based on their own deeds. If sales went down eight per cent last month the initial answer is a call like “we need to sell more”. Another very common way of facing revenue problems is to cut costs: “we need to move our production to another country” could be one thing to do, “we need to force our vendors to lower their prices” is yet another.
That’s all fair and congruent with what industrial economics keep telling us. But there are a couple of things which are very well hidden from the naked eye when we look at things this way:
Most companies are relying on the implicit assumption that their customers would prefer to buy products from their existing product or service range.
Changes in web site traffic are very often attributed to the company’s own efforts to trigger attention: a new product is announced, a sales campaign is done, a loyalty program is started. If figures go down, they are brought back up by promotional efforts.
The data that is looked at usually comes from the company’s internal data sources: how many newsletters have we sent out (and how many of them have been opened)? How many times was the link to the special offer clicked (and how often did this lead to a purchase)?
What you can see straight from looking at Web Analytics data is: How much, how many, how long, how often? All with regard to the company’s own operations within a market that is considered to be stable or rather influenced by the company’s own decisions.
Other things could have happened in the market place as well: At the same time a new competitor may have released a similar product at a lower price (or with better quality), the purchase power of the customers may have decreased heavily, or the roster of decisionmakers may have changed entirely. In all cases a decline in customer interest is fought with the traditional methods of advertising and attention grabbing.
But what if the underlying assumption about the value perception of products in a market was attributed wrongly in the first place?
Taking the example of the three mentioned companies from the introduction and cutting a long story short we can as well concede that the reasons for these companies to strive after WW II was rather grounded in the rapidly increased space in newly erected houses with their cupboards, wall systems and clothes rails.
What if the huge demand for porcelain, model railroad boxes and affordable clothing was simply a compensation effect for the horror vacui caused by the war?
A more recent example for what might be considered a similar misattribution is the ever-repeated lamento of the music industry: their biggest pike in sales was reached in the year of 1999 (although we kept hearing the “Home taping is killing music” meme since the Eighties).
It would be interesting to see how much of the 16.4 billions of Dollars spent for CDs in 1999 [data coming from an article in the New York Times, January 2010] went to music released in that year – and how much of it was spent for filling up the back-catalog with re-released music from the 60s and 70s.
What we can concede for companies or a whole industry is a increasing danger for misinterpreting the core reasons for people’s interest in their products. Long-term success in markets becomes more and more unlikely, as the environmental factors become more volatile. It all boils down to a companies’ ability or inability to continuously refine their sensorium for their environment while keeping up their core operations in a distinct manner.
What if: a company would have no means to draw a distinction between a call for change which they cannot afford to ignore from a call for change which directly threatens their core constitution?
Here’s my second hypothesis:
2. Modern companies are no longer run by people – they are run by business administration principles
There are a couple of constituents for companies regardless of their product portfolio: they need to get in more revenue than they spend on their operations, they need to keep in mind that from the perspective of saturated markets they need a particular strategy for tackling their markets, and they need to be able to incorporate change with regard to their product portfolio as well as with regard to the organization of their own deeds and observations.
On a strategic level, the servo mechanism of applied business logic leaves two distinct ways for running a business: either a company strives for quality leadership or for price leadership.
Both concepts, when applied in their purest form, are calling for different tactics in tackling their respective market segments and customers. Sociologists may call this the application of different lead distinctions, but it boils down to the idea of focusing on “effectiveness” or on “efficiency”, contrasted with the ever increasing importance of standardization and productization.
The idea of effectiveness centers around the principle of aiming to realize the best possible outcome with the input efforts at hand. For service-oriented businesses like banks this could be a “personal financial adviser”, which is a personally known bank employee who is taking care of all financial aspects for a selected group of customers, based on excellent familiarity with each of his client’s life situations.
Efficiency, on the other hand, is centered on optimizing the input relation side of things, trying to offer a reasonable service at affordable costs (for the company). An efficient bank service may come in the form of a call center number, where all of the bank’s customers’ data and personal information are stored in computer systems. Call center agents can retrieve that information in nearly real-time and put back notes on covered topics, mentioned products, information needs, and customers’ potential interest in any of the banks’ products.
Particularly the growing interest in products (those who run a business seem to be genuinely more interested in products than their customers) seems to put a favour on the efficiency side of business systems.
Staying in the example of the “services” portfolio we can see an increase in the amount of call centers, and, in particular, we see an increasing utilization of call center agents as sales persons.
The reason behind is simple: Rephrasing customer inquiries as a request for information about a given product kills two birds with one stone: the idea of customer support and counseling is cross-referenced with directing customers’ input towards refining an existing standardized product portfolio.
This can be easily transferred into a value expectation for more sales (with less costs: the call center could be situated in India), as direct feedback about the relevance of an offering is collected from the marketplace and the communication about the products can be adapted in a way that the value perception of a product can be separated from its factual value.
What if: the idea of entrepreneurship is nowadays decomposing into a sequence of decisionmaking processes which are fueled by business administration principles in uncertain market situations and no longer centered around striving for excellence in addressing customer needs?
The importance of Web Analytics for understanding companies and business administration principles
Admitted: the outlined hypotheses regarding companies and businesses are a bit like a xylograph: a lot of black and white, and no grey area in between. Still following the “contrast folio” idea we can re-formulate the original hypotheses as a need for developing and re-evaluating a sensorium for the environment a company is operating in, and we can see a need for optimizing the way how a company communicates the value proposition of their products.
Web Analytics tools are primarily means for counting and measuring deeds on companies’ web sites. The predominant data type is that of clickstream data where discrete points can be distinguished: a visit on a web site has a particular entry point, occasionally some more clicks during the session, and a distinct exit point.
When working with companies trying to utilize Web Analytics over the past years, I noticed a tendency towards looking at the inside of the data. It may be due to misleading promises from tool vendors, it may be due to a general misconception in effective complexity reduction, but most Managers I have met seem to expect analysts to reconstruct their company’s business realm entirely from click stream data on their own site.
To put it bluntly: Web Analytics data is directing the attention towards the inner operations of a company. The reason for that is mostly that their web sites are depicting the inner state of the organization rather than orienting towards the expression of their customer’s desires. The question they are explicitly asking is: “What is happening on our web site?”
The answer to this question is commonly said to be addressed by generating reports or dashboards. These reports mostly come with a comparative dimension: this month’s figures vs. last month’s figures.
Contrasting this statement with the first hypothesis (“companies resist change”) boils down to the general question: “What kind of change is acceptable?”
It’s of course positive change which indicates a growth in visitor, visit and page view figures, as all of this indicates opportunities for business growth.
Unacceptable changes would be: declines in volumes, transforming into smaller likelihood for business and sales.
An ongoing period-to-period increase in the measured dimensions of page views, visits, and visitors, seems to be the conditio sine qua non for company web sites in general. I tend to call this the “Analytics 0.9” perspective as it allows statements like “Mick is taller than Dave. Dave is taller than Richard.” Analytical conclusions to be derived from this are of the nature: “Mick is taller than Richard”. Nothing wrong with it, but doesn’t lead far.
Looking at Web Analytics data in the light of the second hypothesis (“companies are run by business administration principles”) gives a much more elaborated view on things: “How well does our web site support our business objectives?” is a question I would label as “Analytics 1.0” question, as the comparison pattern here is no longer absolute figures or general trends but this question addresses the company’s value creation principle for their web site.
Admitted: contrasting a “What…?” question with a “How…?” question may be considered a questionable attempt, but there’s a very valid point in doing so, as it depicts a paradigm shift in Web Analytics:
The initial territorial acquisition of the Internet as a marketing channel created an additional consumer touchpoint. As the “Internet population” was increasing, the visit volumes on company web sites were expected to raise as well. This is the very core of the month-to-month comparison of absolute figures within reports – and this used to be a constant source of good news for years.
Putting the web site developments into context with the business objectives is a genuinely modern conceptualization of Web Analytics. The novelty lies in establishing a weighted reflection value for the data at hand. This is infusing the tool character with both the possibility and need for more elaborate and in-depth assessment.
An example: for an e-commerce site you can calculate two easy performance indicators: the “average order value”, and the “average visit value” (instead of calculating the second one you could as well pull up the “visit conversion rate” as a reference figure).
The core point here is that both calculations are complementing each other: if only a small fraction of visits results in an order the average visit value will be low (as is the conversion rate); if the orders received are predominantly low volume orders the average visit value will be low, too – although the conversion rate figure (another very popular performance indicator) is still on an acceptable level.
These two calculations are only loosely coupled – but together they enable a powerful (yet arbitrary) reference horizon for interpreting the change in the data from period to period.
The shift from “What?” questions to “How?” questions requires shifting the perspective from absolute figures to indicative figures. “How” presumes a carefully constructed reference point which can be equally well centered within the business logic as in any other relevant domain for operating a web site (like: the user experience, the user engagement index, the site reach index,…).
These observation domains are not built into the architectural logic of a web service, nor are they reflecting any hard-wired relationship between data and their references. By giving up a tight coupling (to captured data & month-to-month reporting) in favour of a loose coupling (to attribution & analysis) the communication and interpretation of the web site performance gains degrees of freedom – and only creates the space for deriving actionable insights.
When shifting from “What?” to “How?” the applied reference layer becomes partly invisible – this is the reason why technicians often take an entirely different perspective on a web service than marketing representatives: all participants in the discussions are regarding the structure of their own observations as invariable, but they don’t communicate their underlying assumptions and starting points.
This creates a paradoxical situation which can only be mended on yet another level of observation and discussion: as soon as the “Why?” dimension of data interpretation is added, the underlying attribution logic again gains some visibility. (This topic will be covered in part 2 of this article.)
The discussions about Web Analytics, its contributions and limitations is a bit like a picture puzzle: depending on how you look at it you see different motives.
For sure having something measured is better than having nothing measured. A comparison over time can at least generate some insight into trends (The “What?” level). On this particular level the technical details of the measurement are not contributing much bias (given that a statistically significant volume is achieved).
When interpreting Web Analytics data and results the consideration of business objectives is a major point, but fuels a lot of misunderstandings, too. Most web sites do have a multitude of purposes and business goals. Each and every goal can be validated against different data sources and performance calculations on the “How?” level. As said before: the underlying business objectives aren’t necessarily visible during this discussion any more.
Analysts are thus well advised to include their perception of the underlying business objectives of any given site or service to their audience. Other than that they may miss out on the “recommendation” part of the analysis and may obfuscate any possibility to continue discussions beyond pure data analysis.
Commonly three different purposes for web sites are named from a business perspective: generating leads, increase sales, decrease costs. These purposes are usually contrasted with the company’s core business administration principles of efficiency, effectiveness, and targeting.
Looking at both dimensions together establishes the possible tactical approaches for further web site development:
Generating leads efficiently could mean to make it easy for visitors to leave their contact details behind and to make them specify the type of information they are interested in getting. To increase effectiveness, related information about their position, company turnover, and buying horizon, would be beneficial. Collecting and utilizing this information wisely ideally results in targeting the communication and offering through company representatives towards the specific request.
In order to increase sales the applied efficiency principle requires to have the path-to-purchase as short as possible; an increase in effectiveness usually culminates in the call for doing up-selling and cross-selling; better targeting means that you try to get an audience to your web site which is ready and willing to buy.
The purpose of decreasing costs usually has two aspects to it: lowering service costs through self-help services (for Support sites, mainly), and lowering operational costs for the web site itself by outsourcing hosting, content creation, and maintenance.
The degree of success or failure of any of these actions can be measured with Web Analytics tools along the lines. That is the utter core value proposition behind all the tools: you measure, you change something, you measure again.
This all sounds very easy and logical. And it seems possible for all objectives to cater for two things at a time: to combine aspects of efficiency and effectiveness, to have the costs cut and the outcomes improved, to harvest valuable insights while re-shaping the performative edge of the service.
Continue reading the second part of this article: on users