Skip to Content


That’s a quote from a recentZDNET blog post by Brian Sommer and in it he suggests that the notion of an accounting centric ERP system is fundamentally flawed, that the social media, cloud computing and external event driven systems somehow make a major difference to the way ERP needs to be for everyone. He didn’t say ‘everyone’ but he suggested it.

Gas PumpsManufacturers for example will argue that they are influenced by external events like commodity prices, promise dates and availability but often they engage in smoothing, they have always engaged in smoothing, it was the only way to ensure business continuity and they work on statistical margins to maintain that.

Oil & Petrol as an example

MRP doesn’t necessarily need to know that the commodity price of item X has gone through the roof, but procurement, costing and sales will want to know that when they do their next major assessment or negotiation in order to ensure that they aren’t blindsided and potentially run out or sell newly made stuff at a loss.

It is a part of ERP, but is really a part of the transactional processing part of ERP? In these days of accusations of price gouging by the petrochemical industry, gas-pump pricing is seen to be directly impacted by the price of a barrel of oil or the socio-political strife in a given oil producing country but how much of that retail price adjustment is directly associated with costs and how much is associated with opportunistic price adjustment on existing inventory holdings as part of a futures market manipulation? Has, for example, the price of motor oil, increased by the same factor as gas at the pump? Aren’t they both derived from the same source?

This idea of the classic accounting centric ERP has been buffered by criticism for many years now, primarily by pundits of new generation technologies and those producers of software trying to break into the marketable software world with new and innovative ways to do stuff, mobile solutions and business analytics seem to have become very focal of late.

HANA was critiqued in the article for being “same-old, same-old” but with a different engine. Perhaps a bit like a sleek lined classic car with a modern high performance engine. To assume that HANA is enough is of course a mistake, to assume that HANA is in the wrong direction is an even worse testament to a lack of fundamental understanding of the end-to-end dependency that business has, on proper accounting, proper control and fiduciary responsibility for which ERP’s like SAP have been renowned.

Will you really deeply discount or increase the price of a  line of products in response to a flurry of tweets about them?

Business is not “in business” for any reason other than to make money, and money is measured in a multitude of ways. Sometimes the money is accounted for well, and in other instances…less well or rather, less transparently, unless your business model is about playing futures markets, what’s the relevance?

Simplistically it is what we sell stuff for less what it costs for that stuff. What is left over, we either give to the owners or reinvest in the business to make it bigger; we might do a combination of giving back or reinvesting. If it weren’t for the accounting part, we’d have no idea whether we had anything to give-back/reinvest or not. There are some alternative ideas on how accounting should be done but essentially no-one in their right mind ever disputes the logic of having a balance sheet and a general ledger.The idea of having a constantly changing cost model in response to external forces is interesting but it would be a little unnerving for investors or prospective investors if they felt your business was so susceptible to market fickleness. I’d be interested to know exactly which businesses that are heavy ERP users really use this model – financial instruments and short term real estate comes to mind, but are there others?

The more diligent amongst us likely do checking account reconciliations every month. Determining how much we should have in our accounts (ledgers) relative to what we earn and what we spend determines a part of our net worth enumerated in monetary terms. By default we have all bought into the idea of accounting and even if it wasn’t measured in dollars and cents and was instead measured in number of livestock or number of bags of wheat, that  is still accounting.

Is HANA about the failure of business analytics snapshots?

It struck me during SapphireNow 2011 that the reason why HANA is so appealing is that the historical analysis methods that have been touted for at least the last ten years to ERP customers have partially failed. The reason that they have partially failed is that once again we have reached a situation where something was engineered in the software world to cope with inadequacies in the hardware world. Disk caching immediately comes to mind for me…

One of the reasons why we migrated data off to data marts, analytics databases and the like, was the fact that running complex analytics against the live ERP data negatively impacted the rest of the users. Users who werew working on the system were shoveling the data in that we wanted to use for the analytics.

We created complicated extractor solutions that transferred that data to a separate database (often in aggregate) and we did this periodically, again to alleviate the burden on the OLTP. We suggested the implementation of rigorous archiving strategies for data objects that were now no longer required to be immediately accessible online and we built entire infrastructure just for offline analytics.

What we still struggled with though, was what I refer to as tactical reporting needs. These were daily needs that couldn’t be addressed by the delta load to the offline analytics database and which required us to still run reports and extracts against the live ERP. Extracts for the data that had been shoveled in since the last synch with the analytics database.

For those special needs that identified a requirement of granularity in the transactional data that wasn’t available in the analytics database and for the long-view that showed all the activity for a given account, customer, material, vendor or business unit we developed queries, reports and special jobs that ran the risk of degrading ERP performance.

We did it because we needed them for effective operations.

With the passage of time, we have become increasingly dependent on these reports, so much so, that they have become the focus as an area that we want to optimize and improve performance. We want to do this without negatively impacting the rest of the system landscape and operational user-base. HANA perhaps will address some of this.  For now, hardware appears to have caught up a little with where we want to be. Computing memory is now cheap enough that large quantities can be heaped upon our systems that give us more flexibility in being able to faster and more complex analytics with live as opposed to snapshot data being manipulated. HANA will of course leverage that.

We will always need the snapshot data. The need doesn’t go away, but now we have the promise of more timely, more accurate and more relevant reporting capabilities for tactical action. There will need to be some application rewrites in some areas but to assume that we have to completely rewrite all of ERP assumes that much of what we have is fundamentally flawed or broken and that simply is not so.

The data entry tsunami continues

People still need to enter timesheets, invoices still need to be processed, orders placed and taken and assets and liabilities accounted for and those areas do not necessarily need to be fundamentally changed to accommodate the tablet computer or smart phone user. What we have to accept is that the data volumes are growing, unabated and that our ability to report in summary or with granularity needs to be made possible, we’re not content with just integers and text either, sometimes we also want images!  The notion of processing hundreds of schedule lines on a tablet PC is ludicrous to even consider. Display or inquire… perhaps.  The extension argument that you would build complex system integration like EDI or PI/XI  for every multi-line-transaction to alleviate back-office pain is an equally flawed idea, the creation and maintenance cost is hard to justify especially if it comes from IT’s budget.

One visitor to the Winshuttle booth at SapphireNow talked about the fact that they produce thousands of manpower requests to vendors every day, that these cannot be done based on contract orders because they are for a limited pool of discrete resources in the market but that the rates and margins can have significant cost vs. profitability impact. The ability to extract all the purchase orders in gritty granular detail for hundreds of business units ‘on demand’ is something that current data warehouse and data mart approaches aren’t usually configured to support.

In fact, changing them to accommodate this need is probably also not the right thing to do, given that there is still a requirement for aggregated trend analysis and filling the analytics database with superfluous granularity would just slow down the routine analysis.  So they have a need to do real-time trended analytics against the live ERP data. They’re struggling to find solutions that are a good fit either from the SAP home stable or from the SAP Ecohub  partners but they are hoping that HANA will help them in getting where they want to be. In the end they may have to go for a hybrid approach but more importantly, every day that they don’t address this need, they leak margin as they overpay for resources that they could better negotiate rates on.

To report this post you need to login first.

Be the first to leave a comment

You must be Logged on to comment or reply to a post.

Leave a Reply