Skip to Content
Author's profile photo Thorsten Franz

BW on HANA: The Value, Seriously

We’ve heard it before (and we experience it where I work) that with BW on HANA, you get significantly faster, more responsive query execution and you can use more fine-grained data. But what does this really mean – where does the real value come from?

Business insights: up from 0 to 100%

Please look at it this way: If every navigation step takes 30 to 40 seconds, then how likely is it that end users will derive actual insights from the data they have in their BW?

Imagine the following. I’m handing you a complicated 3D object. I’m not sure what’s wrong with it – perhaps it’s broken. Either way, you are supposed to examine it carefully so that you can make sense of it, understand the structure and perhaps find out what’s wrong and how to fix it. You can hold it in your hands, rotate it, and look at it from every angle. But there is one restriction: Every time you want to turn the object in your hands ever so slightly to change the angle from which you look at it, you have to wait three minutes. How far are you going to get with all these frustrating interruptions, and how long will it take you to find a more worthwhile endeavour to invest your time and attention into?

My experience is that having a slow, unresponsive BI system gives you as much value as if you would send all of your data to /dev/null. If end users don’t find the response times good enough for an uninterrupted, question and answer like workflow, they won’t use the system, and it is either completely useless or leads a half-life as a platform for data extraction, transformation, aggregation, and forwarding to a separate platform that allows users to actually work with the data.

So using HANA to speed up query response time adds more value than most people might think. If a number of queries that used to run five minutes each are sped up by factors of 10 to 20, that doesn’t mean that a few folks can go home a bit earlier or make themselves useful in other ways during the maybe ten minutes they might have saved. Forget that way of looking at the value. The added value is that they do something they haven’t actually done before, at least not as extensively: work interactively with the data, look at it from many angles, and derive actual insights that lead to better business decisions.

Observe and adapt, anticipate and lead!

With BW on HANA, you can simplify your business content and data models. Frequently, these are more tailored around the performance bottlenecks than the actual functionality or business logic. With BW on HANA, you have the chance to drastically simplify your data model, maybe do away with the majority of your artifacts (such as DSOs and the programming logic between them). This is an initial investment, but pays off quickly because of two effects.

  1. It reduces your maintenance cost.
  2. You gain increased agility, so you can create new business content that is relevant to the business faster and tailor it more closely to the needs of the business.

The latter effect is more important if you ask me, because we live in a time of rapidly changing markets and economies, where the business has to be smart and observant in order to survive and have a competitive edge when the next big transformation in the market takes place. Need to observe and adapt, anticipate and leads. That’s only possible with business intelligence solutions that can evolve at least as quickly as the world around.

Other effects, less fundamental but not to be neglected

  • Data staging might become faster, resulting in earlier availability of information to end users.
  • From a buying and TCO perspective, this might be an opportunity to throw out an existing BWA and simplify your system landscape.

Prepare for the great convergence of OLAP and OLTP

Here’s another fundamental one. Looking at customer system landscapes, I can understand what keeps CIOs up at night. Here’s an ERP system, there’s a CRM system. Data goes into a BW system. Then it goes out of that BW system into another system, this time with a big database where it is merged with data from non-SAP systems. Then it is propelled forward into yet another system that has great query tools. If you speak a lot with customers in the BI field, you might find that this is quite a normal setup.

But think for a moment about how expensive it is to maintain all those different systems that have different life cycles, to figure out how to synchronize and harmonize their needs to be patched and upgraded, to have all the relevant skills for administration, content creation, data transformation, troubleshooting and all that in your company. Think about what a waste of time it is when the data has to go in and out of four different systems before it can be used for reporting. Think about how difficult, slow, error-prone and expensive it is to change something in such a setup, when four systems are affected.

Thinking of all this, I can see the huge strategic value of the OLAP-OLTP convergence that becomes possible with HANA. The vision of one platform where the same data can be used for transactional processing, embedded analytics and standalone analytics is so very compelling because it simplifies the big picture, removing many elements that drive costs without driving value, that slow us down like an iron ball tied to our ankles, and that ultimately stand in the way of IT to becoming the decisive factor that wins the game.

So yes, the OLAP-OLTP convergence is not to be missed, and adopting BW on HANA is the logical first step on that journey. As soon as you have a HANA box, other steps can follow immediately – but let’s discuss those in a separate blog post.

Assigned Tags

      You must be Logged on to comment or reply to a post.
      Author's profile photo Tammy Powlas
      Tammy Powlas

      Very well done.  It reminds me of this video where Harald Reiter interviewed Udo Patzelt of your same company and he spoke of OLTP/OLAP convergence as well.

      Author's profile photo Sascha Wenninger
      Sascha Wenninger

      Great post, and totally agree with your position that latency is a killer for use cases beyond canned reports, ie. the high-value stuff. Probably for that same reason is SE16 one of the most popular transactions in most of the ERP systems I've seen - users like getting their data into Excel and trying things without latency or engaging IT... Can't blame them either; it's their data after all 🙂

      Author's profile photo Former Member
      Former Member

      once we discover that we can just navigate one giant pivot table we don't need anybody to design, test, implement, and support any reports for us.

      that's like hiding all the complexity in the world that we are happy to dispense with...ok, i'm dreaming

      Author's profile photo Nigel James
      Nigel James

      Great thoughts Thorsten. This morning I was given an old laptop as a possible replacement for a machine that was having intermittent issues. It was powered by a centrino chip. It was so slow that it took 5 seconds to paint the start menu on pressing the windows key. Needless to say it didnt last very long and I went back to my intermittent error.

      Latency is a real challenge when working with a complex system and HANA certainly removes that from a analytics scenario.

      Thanks again for a great blog.

      Author's profile photo Subhankar Pattanayak
      Subhankar Pattanayak

      Great insights. I strongly believe that with faster technology we can solve, we can have better business solutions.....I had written a similar blog here:

      Author's profile photo Hendrik Brandes
      Hendrik Brandes

      A very well done blog. Think about travelling more by plane 😉

      I think this was one of the most reasons, why customers want to have products like "local In-Memory"-Databases or something else, where they could experience their data - instead of having them painted on a great wall of numbers.

      After SAP has respect, that customers/users want to work with their data and not only show them, I think we will have much more exciting tools in the future like "Visual Intelligence" or "Predictive Analysis". And those do not make any fun if you have no fast access to your data.