Data-Driven Marketing in E-Commerce

Steadily increasing competition, niche marketing which has been driven to the extremes, as well as the basic professionalisation of e-commerce have resulted in an unprecedented level of complexity. This is especially true when it comes to the marketing of products and services. Online retailers need to take data-driven marketing in new directions.

Objectives and planning in data-driven marketing

Data, data and more data

Without the constant analysis and optimisation of the various marketing measures, companies are at an increased risk of running campaigns that ultimately cost more than they generate. The key to successful online marketing lies in continuous data collection, evaluation and optimisation. True to the saying “the numbers don’t lie”, decisions should be made exclusively on the basis of ascertained facts.

Web analysis alone is not enough

E-commerce controlling is the solution. In contrast to web analysis, controlling involves the analysis not only of movement data within the online shop, but of all company-critical data.

From Google AdWords to the complex marketing mix

In the past, only well thought-out SEA campaigns, such as Google AdWords, were able to generate the traffic necessary for business success – as well as the relevant subsequent sales – but this strategy is hardly successful any more today. The costs per click are too high and the competition situation too fierce. Therefore, as in the case of classic offline marketing, a mixture of various different marketing measures must also be developed for e-commerce.

Different objectives in the mix

In doing so, each measure can and should pursue different objectives. The profitable sale of a product does not have to be the exclusive focus of a campaign; further objectives such as the generation of traffic and increasing reach are just as legitimate. Particularly when selling more complex products with a long pre-sales phase, lead generation and considering the customer journey as a whole are paramount.

SEA and SEO are not a panacea

In practice, it is all too common for marketing measures to lack a well-thought out and comprehensive strategy. Often, companies cry out for SEA and SEO without much thought as to the effectiveness of these measures in the achievement of the desired objectives. It is time to say goodbye to the notion that SEO and SEA are the ultimate marketing miracle weapons that can solve all traffic, sales and lead problems in one fell swoop using good campaigns.

Comprehensive marketing strategy as a basis

The development of effective marketing strategies is primarily about identifying the “right” marketing measures as well as the most sensible objective in relation to the respective marketing channel. It makes no sense, for example, to use marketing channels in which the target group is not active, or to use a channel suited to the generation of contact data but define sales revenue as the objective. Bearing in mind the intended target group, framework conditions such as pricing and the product range, as well as the available budget, it is advisable to compile a list of marketing measures and then validate which of these are relevant to the marketing strategy.

Data-driven marketing – Selection of objectives

Next, the objectives must be defined. These must above all be measurable, for example the sales revenue to be achieved, or the number of contacts to be generated. As mentioned in the introduction, however, the profitable sale of a product does not always have to be a priority. Particularly in the case of traders who are just starting out in e-commerce, the primary objective is to increase their reach. The online shop must make a name for itself and build up a broad customer base over the long term. The actual monetarisation takes place in the second step. In this approach, a sale to a customer may generate a negative profit, provided the likelihood of a second, profitable purchase is high.

Reach with Amazon and Ebay

Multi-channel marketing is particularly suitable for increasing reach. The sale of a product portfolio on Amazon or eBay can be effective, even if these marketplaces are often unable to generate any appreciable profits due to price transparency and sales promotions. Both platforms, however, are excellent for attracting customers and increasing awareness. It is possible to quickly build up a broad customer base, which should then be dealt with in the after-sales department.

Controlling and evaluation of data-driven marketing campaigns

Once the strategy and objective have been defined, the structure of the controlling system, the analysis technology and the optimisation cycles can be addressed.

Identification of Key Performance Indicators (KPIs)

In addition to the definition of objectives for each marketing campaign, key performance indicators (KPIs) must also be defined. The evaluation of success within the framework of e-commerce controlling ultimately functions only on the basis of these indicators. If these are only vague or have not been defined at all, no statement can be made regarding the success or failure of the campaign.

Examples for KPI sets: Google Shopping

If, for example, products are marketed using price search engines such as Google Shopping, the KPIs might look like this:

  • Number of clicks
  • Average length of visit in minutes
  • Average order revenue
  • Revenue (total)
  • Production costs (total)
  • Advertising costs (per click)
  • Advertising costs (total)

These KPIs allow you to determine whether the campaign is cost-effective and how much traffic it has helped to generate.

Examples for KPI sets: Adwords campaign

In contrast, a Google AdWords campaign might use the indicators “Average Order Revenue” and “Revenue (total)” when the objective is solely to increase traffic and generate leads. This definition of objectives would in turn result in new KPIs, such as volume of contact data generated. The definition of success-critical KPIs therefore depends primarily on the marketing channel selected and on the objective assigned to that channel.

Determination and evaluation of the obtained data

The method of data collection must then be devised and technically implemented. The central question here is always which system can provide the required KPI and when.
In practice, it is only very rarely that a single system is able to gather and process everything necessary for data-driven online marketing. Generally, a combination of the systems which form the IT infrastructure is used. The following tools were selected for the Google Shopping campaign, for example:

  • Web analysis software, such as the econda Shop Monitor, focuses primarily on the transaction data that is generated within the online shop.
  • The Magento e-commerce platform enables data management in relation to orders.
  • The ERP system, on the other hand, contains company-internal information on purchasing prices and logistics costs.
  • The costs of Google Shopping marketing campaigns are borne by Google.

Contextualisation, centralisation and visualisation of the data

For a comprehensive and valid overview, it is imperative to put all data into context. If this is not done, there is a risk that the marketing campaign will be assessed only on the basis of the incurred costs and the available budget, but not in terms of the sales revenue generated. Conversely, the pure transaction data from the online shop provides interesting information such as the conversion rate, but no information about the resulting marketing costs.

A business intelligence platform makes sense

In order to resolve this problem and to put all data from a wide range of systems into context, a business intelligence solution for Magento shop systems is used: Magento BI. BI solutions prevent the isolated analysis of KPIs – and thus their erroneous interpretation – by connecting different data sources and preparing graphical reports in practical dashboards. This is the only way to really ensure that the efforts aimed at establishing data-driven online marketing are not in vain.

Data-driven marketing – Optimisation and further development

Once all key indicators for each campaign can be properly recorded and analysed, the further development and optimisation of marketing activities can begin. Let’s stick with the example of the Google Shopping campaign: Here, you could test what happens when the marketing budget is increased. Do sales stagnate while costs steadily rise? Or do sales increase faster than expenditure rises? And – looking at the bigger picture – is it more profitable to generate twice as many orders, or does this result in logistics costs increasing exponentially?

Measure against reality

The optimisation and further development of marketing activities is primarily about “testing”. In the tests, only a small number of modifications should be made at any one time, in order to be able to trace any resulting change to the relevant modification.
It is almost impossible to specify in advance which marketing channel will perform using which setup. Tight controlling of the numbers is thus a decisive factor in data-driven online marketing. Through consistent and comprehensive data collection and by focussing on KPIs and measurable objectives, it is possible to make a statement about which marketing activities “work” and which simply burn money.

The contribution to the topic of data-driven marketing is taken from the ‘Future Topics in E-Commerce 2017‘ annual on trends and best practices from the online branch of our partner netz98.

scroll up