When and what to automate – Marketing Automation pragmatically

Picture of Tomasz Karwatka
Tomasz Karwatka

Supervisory Board Member at Divante. Leading industry voice who believes eCommerce can improve our world. Co-founder of Vue Storefront and Open Loyalty, angel investor, and founder of Tech To The Rescue. CoFounder at Catch The Tornado eCommerce Startup Studio

Care to share?

During the Industrial Revolution, a huge steam engine used to be put either on the ground floor or in front of the building and it powered the machines in factories through conveyor belts. It was an unreliable solution. When the electric motor and the electricity itself became popular, they replaced the steam engine. One machine was put in the other’s place. Conveyor belts were still in use.

Only after a long time, someone came up with the idea that many small electric motors could now be used for particular machines in the factory. It allowed to eliminate one of the points of failure (one engine was replaced with many small ones) and to substitute unreliable conveyor belts with electrical wires, which were much easier to install and to maintain.

Why am I writing about this? Technology evolves very fast. Our abilities to adapt technology wisely usually do not keep up with its evolution. It is a repeating pattern.

Let us now move on from factories to marketing, still dealing with automation.

When we ask users of Marketing Automation tools (http://venturebeat.com/why-marketing-automation-fails) what obstructs their work the most, we receive a very interesting answer:

  • those who are satisfied with MA respond that the biggest challenge is to create contents (in big amounts),
  • those who are not satisfied with MA respond that their main problems are integration of systems and a mismatch between their functions and needs.

I put forward the hypothesis, that users unsatisfied with MA simply do not know how to adapt this technology, they have not discovered it yet. They implemented automation without knowing what and why to automate.

Big companies always try to implement automation processes. Thus, many ideas get rejected at the very beginning. One of our customers, who owned an orthopedic shop, had an idea on how to increase the sales: an online chat with a physiotherapist. This chat is very difficult to automate. For this reason, a completely manual control was chosen. It turned out that conversion from the conversation with a physiotherapist to a purchase amounted to 40 per cent, a 20 times better result than the average conversion. This result is so good that automation is not necessary and more consultants can be hired.

This project inspired us to realize a Marketing Auto in the form a step strategy. In this kind of strategy, first we focus on creating hypotheses and discovering what functions well. The automation process does not begin until later.

Step Strategy

According to optionmonster.com, as many as 70 per cent of users who abandon a website, will never come back. Thus, creating hypotheses usually begins with ideas for building a user base. Acquiring a customer is increasingly expensive and thus we believe that creation of a user base is fundamental. A base means for example subscriptions on social media, newsletter signups, leaving the phone number – everything that makes it possible to contact the customer directly and on our terms.

By putting forward more hypotheses we discover for example what makes users sign up for a newsletter. This way, in 6 weeks we increased the e-mail base of our client from the fashion industry by 34 per cent. It also turned out that only the sales from the welcome messages (after signing up for the newsletter) increased the shop’s sales by as much as 14 per cent. People who are signed up for the newsletter buy more – the size of their basket, in comparison to an average basket on the website, is bigger by 17,51 per cent.

What’s next?

After a positive verification of hypotheses, we automate chosen solutions. We usually realize automation using existing Marketing Automation solutions. We apply the existing solutions in order not to involve the IT department on the client’s side.

After a positive verification of the automated procedure, the moment of the integration process arrives.  The integration requires an intervention in all the IT systems of the client. A premature integration (without verified hypotheses and successes in automation) leads directly to a catastrophe (exceeding the budget and time limit, no business effects).

How can this integration look like?

For one of our clients from the electro-technical industry, we realized a project of building and linking of the eCommerce B2B system, CRM system and PIM (information about the product) and the storage system. It allows to realize an automatic and coherent customer service in each channel.
Example: the client visits the online shop and adds products to his/her basket; a system of intelligent recommendations creates propositions of complementary products; when products are still in his/her basket, a call-center consultant calls the client with a proposition of additional products; after a confirmation the consultant can add these products directly to the client’s basket. The client is surprised by this pro-active approach and consulting, he/she still controls the situation and can see a preview of the whole online order.

This system currently generates over $50,000,000 per year. To obtain this effect, an implementation stage of over a dozen of months was necessary. It would be unreasonable to begin from the integration of technologies that have not confirmed their usefulness yet.

Published January 19, 2015