ca. 30 Minuten
nach Wunsch auch mit VideocallA comment on Forian Heinemann's data warehouse podcast on digitalkompakt.de
Rarely have I been so impressed while listening to a podcast as I was with Florian Heinemann's performance on Digital Kompakt on Oct 21st 2016. Why? Quite simply: He speaks to me from the soul. For almost ten years, I've pestered people about how essential data-driven work is to the success of a company, both on a strategic and operational level – and I almost fear they cannot hear it. So it is extremely pleasing when another voice finds the same words! Because I can only agree with what Mr. Heinemann says, his main theses are summarized here, along with some comments from my personal perspective. In the “Here’s how this thesis is reflected in minubo” paragraphs, you also read how Heinemann's theses are reflected in our Commerce Intelligence Suite minubo. Have fun while reading!
Whoever Reaches for the Stars... Needs a Data Warehouse
It's a simple truth: the competition in the digital sector is tough, so only the companies that make the best business decisions will make it – and the best business decisions are made only with a holistic data base, or with data that is easily accessible to the whole company and the resulting insights that come from this. The earlier a company makes these things available, the better the chances on the market and the lower the initial expenditure on the project – my experience is fully in line with what Mr. Heinemann reports. It is not for nothing that digital giants like Amazon and Zalando have devoted entire business sectors to this topic.
Here’s how this thesis is reflected in minubo:
Companies need a standardized, fast-to-implement, reporting and analytics solution that results in data transparency and which allows them to be fast-paced and stay in business. That is why we have developed the minubo Commerce Intelligence Suite: to cover everything from the cloud infrastructure to the data warehouse, to the reporting and analytics front-end, tailored to the specific requirements of digital commerce and, of course, adaptable to individual needs.
Data-Driven Decisions are not a Management Privilege
It is a widespread misunderstanding that reports, analyses, and all kinds of colorful diagrams are a privilege of leadership. On the contrary, while strategic decisions are made primarily on the basis of high-level evaluations generated in the monthly, quarterly and yearly cycle, operational units as well as lower levels of management need performance data on a daily basis to take the right measures based on solid data. Which campaigns need more budget in the short term? Which customers have a high buying probability today? Which top sellers are running out of stock? Here, only those who have the right data are able to do business. On this point, I can only agree with Mr. Heinemann wholeheartedly – because the operative component of the data-driven decision-making culture is still neglected.
Here’s how this thesis is reflected in minubo:
High-level reports are only part of the story and they serve a specific, limited audience in a company. For this reason, we have equipped minubo with all the necessary types of management reports, including timely plannable e-mail dispatch, but we are particularly focused on operationally usable insights from data as well as proactive recommendations for action. More generally: simple and flexible data access for everyone. Only in this way can teams achieve maximum results!
Solid Change Management is Indispensable
If you have already been through the introduction of a new tool, you will be familiar with this situation: You are implementing a great system that is only waiting to generate tons of added value for the company – and then it will be lying unused in the corner (or in the cloud). What you learn from this: It never (!) works without change management. As Mr. Heinemann explained, this means for the introduction of a reporting and analysis solution: employees must be reminded of why they should not only look into their traditional operating systems, but also into the corresponding performance figures or data-generated insights.
Here’s how this thesis is reflected in minubo:
During the introduction of minubo at our customers we have been working with accompanying change management for a long time – even a whole advisory department has been created for this. With use case workshops, training courses and dedicated customer support, we ensure that companies achieve maximum added value with minubo.
Yes, it Must be raw Data Level
Clearly, it would be much easier (and cheaper) to process data in an aggregated way. The problem: What is processed as only aggregated can only be output in aggregate – and this is enough for management reports, but not for operational work. This already begins with the matching of various data points (for example, returns to acquisition campaigns – this is only possible via individual order IDs) and extends to flexible ad-hoc segmentation, for example, for newsletter campaigns or Facebook target audiences (this can only be done on the basis of individual customer records). Just like Florian Heinemann, I would like to underline: Don’t skimp at this point – it would definitely be the wrong!
Here’s how this thesis is reflected in minubo:
In order to meet requirements, such as the above, all minubo data processing is done at the raw data level – so our users can flexibly segment their data and drill down to individual object entities, for example, individual customers, individual Omni-Channel touchpoints or products or product variants. In addition, they have the possibility to access their raw data in real-time – whether it is with Excel or, for example, an email marketing or campaign management tool.
Data Flow and Matching are the Heart of Every Analytics Architecture
Numbers taken by the user from their reporting and analysis frontend should be correct and meaningful – and fully updated punctually at the start of the day. This may sound simple, but it is not at all, because to ensure this, a complex, expensive, technical infrastructure is required, whose assembly can only be accomplished with a lot of expertise, both in terms of technology and content-analytical questions – and finally with a lot of time and money. In particular, this is valid for the usual legacy BI solutions à la MicroStrategy or Business Objects. But also, for projects based on open source technology, as recommended by Mr. Heinemann, effort as well as project duration and (opportunity) costs should not be underestimated.
Here’s how this thesis is reflected in minubo:
In our daily work with customers, we observe that the majority of companies do not have the resources mentioned – neither the necessary expertise, nor the time or money. For this reason, we at minubo provide the technical infrastructure and the data model out-of-the-box, perfectly adapted to the professional requirements of commerce customers and significantly cheaper than alternatives, especially from the field of traditional BI solutions. We can still customize where it does not fit, but the customer starts at 80% instead of at 0% – and gets to the goal at a much lower cost.
The key Question: Buy it or Build it Yourself?
Well, maybe I'm slightly biased when I say: buy it – from minubo! But jokes aside: I am a supplier of a core-standardized solution because I am convinced that it is what companies need. The resourcing question is one reason for this (see above), but also, the fact that we, as an external provider, have built our solution based on our comprehensive practical experience and continually incorporate best practices from the industry. This is very important in comparison to a completely new solution by your own staff who may never have implemented such a project before.
One of the arguments that Mr. Heinemann has for self-construction is the associated learning effect for the team. I feel the same way. But: I think the biggest, or most lasting, learning effect in such a project comes from the mapping of the various data streams into the centralized data model – and this process is not only important for an in-house project, but also an integral part of every minubo project. In this way, employees also come to a better understanding of their own data and data processes.
A further argument that Mr. Heinemann has for self-construction is the complete customization of the data model. My answer is: I do not believe that the development of a "custom" data model in the data warehouse offers the greatest added value – after all, the account plan in the accounting is not reinvented for every company. In my experience, the need for "customization" has mostly come from past BI projects, which always lacked something in the data model, so that "customization" is confused with "completeness" or "flexibility in completing" the data model.
It is clear that "completeness" is also a point of interpretation and certainly some companies will find aspects that are not fulfilled in our data model, but I dare say that all the essentials (and more) are there – and the rest is only a question of time, because on the basis of the existing 80% we are flexible and fast in the expansion of the model. And: If you start at 80% instead of 0%, you have more time to focus (in a data-driven way) on revenue-generating activities.
Finally, a small Sneak Preview: In addition to our regular product development, we are currently working on a complete, systematic, key figure model for commerce organizations, which is bridging from our previous eCommerce key figure model to the omni-channel key figure model – basically the DIN standard for key figures in online and omni-channel trading. This system will then also be publicly accessible, not only through our product, because I believe such a systematic, consistent model is missing on the market – and indeed at most of the companies I have gotten to know. We want to go this step.
By the way, since it is our goal to combine as many expertise as possible in this system, we work on it in cooperation with our customers, but also want to make you a part of it! ! If you are interested in actively shaping our project, please write me an email.
Last but not Least...
... a big thank you to Mr. Heinemann and Mr. Kaczmarek for this great podcast! If there were more opinions like this, many companies would have fewer major problems. To anyone who has not listened to it yet, I can only recommend this: Take the 37 minutes and 12 seconds of time – they are worth it!
What are your thoughts on this? Leave a comment and take part in the discussion!
