|
Brief
At karstadt.de traffic data used to be gathered by internal software. This software was replaced with the Sitestat ASP solution produced by the Dutch web analytics company Nedstat. The change enables more flexible use of data gathering and evaluation and has cost advantages compared with the development options for the internal solution.
Strategy
Measuring the effectiveness of the marketing channels (shop)
Whereas the e-mail, banner or SEM marketing activities used to be steered and measured entirely by the supporting agencies, these data must now be merged in a common pool in order to follow the effectiveness of investment to the end.
Integrating transaction data/shipping information (DiVA)
The target of the online shop does not stop at visiting the shop and not even at ordering an article, but only when the order has not been cancelled or returned later. At this level too, data from the newly implemented DiVA shipment handling system must be incorporated in the marketing controlling.
Transfer to bookkeeping (SAP)
At the end of the processing, the data is supplied automatically to the bookkeeping system of karstadt.de. The graphic shows how the data for various process steps can be monitored. Two essential results for optimising performance are marketing controlling and vertical controlling. In the case of marketing controlling, the indicators are set at what weighs most at the level of the individual marketing tools. In the case of vertical controlling, however, the focus is at the level of the different product ranges.
Execution
Communication and employees
Introducing and using the analysis system not only demands a strategic system for establishing the total effectiveness of online sales and technical know-how for conversion but also the implementation of communication processes, linking of the measured results and decisions to translate the results of the information system into actions.
Alongside the usual information channels – e-mail reports and web interface, for example – the hourly measurements of the controlling system are also displayed on a large screen at the offices of karstadt.de. This sort of presentation aptly symbolises the relevance of this information to daily work. More specifically responsible for the controlling is a department with six employees, two of whom focus primarily on drawing up special reports for assessing marketing campaigns.
Target levels and steps for optimisation
At karstadt.de, four target levels are entered for ranking the results in a framework:
Target level 1: range, traffic (visits, page impression, visitors, unique audience)
Target level 2: visit length (visit length, PIs/visit, vertical visits)
Target level 3: transaction (order, newsletter lead, participation in competition, information, website subsidiary)
Target level 4: efficiency (cost-turnover ratio of the advertising investment with sales, including postclicks, CPC)
Based on this ranking, targets for every level can be defined and checked against results. This is done in four steps:
Quantifiable targets = plan
Realistic, feasible monthly targets are set for every marketing campaign and for the verticals of the online offering as a whole.
Permanent measurement = investigation of actual situation
Every campaign is measured as to its effectiveness. The accuracy of measuring the user’s behaviour extends from the origin (newsletter, for example) through his or her behaviour on the website up to and including the purchase and the completion of the order. In this, the cost-turnover ratio occupies a central place. Here the cost of the marketing campaign is related to the turnover generated by it. Finally, turnover is not taken to be the gross turnover of the website but the net turnover after deducting counter-entries and returns.
Deviation analysis = difference between the desired and actual situations
Regular checks are performed to find out whether the targets have been achieved.
Deductions from campaigns to fill the gap
The marketing campaigns are adjusted, drastically changed or reinforced depending on the results of the comparison.
Results
Introducing such an integrated controlling system requires considerable financial and organisational effort. Interestingly, the solution for karstadt.de did not lie in an entirely new solution but largely in the integration of data pools that were already available. In the case of karstadt.de, conversion took more than four months in total. The results, measured on the basis of cost-turnover ratio, clearly show how effectively a productive controlling system can be: for example, an improvement of 40 percent was already achieved in 2006 followed by a further 30 percent optimisation in 2007. This translates into profits from optimisation running to seven figures.
|