Why did ERP Die?
ERP solutions represented a watershed achievement in Industrial Age IT. They were designed when IT was constrained. These systems captured accounting and other internal/operational events and then pushed them from one sub-ledger or subsystem to another. They helped far-flung and disjointed firms finally co-ordinate activities. They brought a measure of efficiency to countless firms and the results were even apparent in the GDP statistics of countries globally.
Yet, these products were always doomed to obsolescence. Moore’s Law advancements in memory, throughput, processor speed, telecommunications, storage density and so on were always going to spell doom. Seriously, how many 30 plus year old technologies are still with us today? Who among you still uses land-line telephones, dialup modems, zip drives, floppy disks, fax machines?
ERP, as we’ve known it, was always fated to die – it just did a better job of finding ways to extend its shelf life.
Those extensions are no longer working. The solution must be rethought, reimagined and rebuilt anew. ERP must be made relevant again. It’s going to be an expensive undertaking for any software firm that has the stomach to bite the bullet.
The Fatal Blows
There are many reasons why ERP is dying. It’s the sclerotic effect of a number of worsening woes that formed both its fate and identifies the seeds of where the new ERP solutions must go. In no particular order, here are some of the medical examiner’s findings:
Industrial versus Digital Economy
A radically different economy and kind of business has exploded onto the scene the last decade or so. Virtually every reputable consultancy out there is trying to help clients of every stripe, industry or size become a digitally reinvented firm.
Old-line retailers have got the religion and offer omni-channel methods for sales while also using big data feeds of weather forecast data to schedule store workers.
Turbine engine firms have embraced additive manufacturing/3-D printing for as much as 40% of parts production.
Car insurers are grabbing patents to put sensors in your car, your steering wheel, and more. They don’t necessarily need the data to sell you auto insurance rather they see a means to sell your road-rage induced high blood pressure, high heart rate and off-color language to health insurers and others while also automagically recalibrating your insurance premiums.
In the digital economy, data is often more important than plants, factories and inventory. The obsession with tangible assets is being replaced by a focus on data and intangibles. ERP wasn’t designed for the latter.
Don’t believe me? Uber, which owns no cabs, is worth over $50 billion. Airbnb, which owns no hotels, is worth more than Marriott. Facebook is worth more than Walmart and Walmart has thousands of stores, distribution centers, truck fleets, hundreds of thousands of employees and billions tied up in inventory. Data is king in the digital era – not fixed assets.
If you look at an ERP system, it’s fixated on plant, equipment, inventory, cost accounting, and other artifacts of the Industrial Age.
Readers must realize that when one era falls, it rarely bounces back. When the Agrarian Age was transformed by modern farming methods, the U.S. saw farm employment fall from a whopping 92% of the economy to around 2% today. Manufacturing, a mainstay sector for ERP vendors is now down to just 12% or so of the U.S. economy.[i] This is why so many newer or re-inventing ERP vendors are focused on service, retail, distribution and other sectors. They’re not that focused on Industrial Age firms.
ERP was never designed for the digital era. And that’s the next death blow.
The Digital Firm uses more than internal transaction data.
Digital firms use all kinds of data --- they are not exclusively using accounting or business transaction data. Some argue that digital firms are the key users of non-traditional data types. That will not prove accurate as more firms become digital. The result is that ERP data is no longer the traditional or only source for business insight.
Digital firms are grabbing massive amounts of information from sensors in their products and assets. They are acquiring third party datasets from a variety of business partners, retailers, data brokers and others. They use both structured and unstructured data. They use data that is hard to validate like social sentiment information, product reviews/rankings/referrals, etc. and data that comes in massive quantities (think Facebook posts, tweets, etc.). They’re using photos, videos and even parsing comments to separate real feedback from snark and sarcasm.
The digital firm’s lifeblood is data that is often huge, lumpy, constantly evolving, mixed mode and sometimes questionable. Traditional data, something that which originally fit on 80-column punch cards, is being supplemented with comments that contain 140 characters, pictures of cute kittens and compelling video.
Everything about an older ERP system was designed to capture the most important INTERNAL business events and efficiently process them. There’s no place in these systems to store all of the new dimensions of data and data types let alone the myriad and changing external datasets.
Those old ERP systems were optimized for on-premises computers using sub-ledgers, relational databases and local persistent storage. The data model for these older ERP solutions never envisioned the digital world today. Trying to retrofit the digital firm into an old ERP solution is like trying to make a Tesla electric car run on diesel. It’s not commercially viable to do so.
If we’re going to replace ERP, then what must the new solutions contain?
One App/Function/Process at a time
ERP vendors must reimagine every niche and every application in their portfolio. While the current rage is to re-work an ERP’s user experience/user interface, this effort is only putting new makeup on a tired old dog. Re-imagination of the processes to work in the digital era will challenge the concept of the application altogether.
To begin, let’s look at how and where "Big Data" fits into modern, reimagined processes. I recently heard one very large, very famous CRM vendor describe how their firm can grab big data, feed it into a big data processing tool and then identify a specific action item that a specific salesperson can act upon. This is a problem.
First, this vendor is using a big data source as feed stock to a funnel-shaped process. The big data is crunched, analyzed and assessed via one tool before sending a single-event/action item to a single worker. Big data wasn’t and doesn’t work in a way that only one actionable item will be the result of the analysis.
Within one dataset, one might identify hundreds, thousands or millions of action items varying from the repetitive to the unique. The vendor in question has built a funnel mechanism as it is an easy way to push one big data insight into a CRM designed for processing one business event at a time.
The old CRM data model was not designed for big data and neither were the applications within it. The old CRM application was, like many ERP products, designed for constrained IT resources and for processing one transaction at a time. This app was NOT designed for the digital and big data era.
Every application software vendor has to look at Big Data as something that impacts not just the beginning of a process but every step of the process. This concept alone has huge ramifications to software companies.
When Big Data can be used at the top of a process for example in marketing funnels, in supplier identification or in passive job seeker identification, it speeds up and automates several steps in those processes. But Big Data can also be used in the middle of processes for example to detect potential fraud activity, to better plan/forecast. It is likely that Big Data can be used at the end of processes as well. Old process designs never considered the use of big data and certainly not at every step of the process. New process designs identify new data types, new data elements, new filters, new algorithms, new action items.
Application vendors must create all-new processes as some processes could not exist in the pre-Digital era. For example, a lot of HR departments are pondering how to measure their recruiting or employment brand. They can peruse what people are saying about their firm on Glassdoor.com --- albeit with caveats. They can also continuously monitor social sentiment about their firm using tools like GNIP and DataSift as part of their HR toolset. Those products permit scanning of social media data streams. Employers could have dashboards showing how their employment brand is trending over time --- but --- HR software vendors will have to build these apps on top of the stream data, design the processes and procure new data sources.
Fix the missing green processes
To illustrate the appalling inertia in ERP software, just look at the almost non-existent progress made in understanding Green technology in the Industrial Age. ERP vendors have known about the concepts of carbon offsets, carbon footprints, etc. for more than a decade. But, did ERP vendors do much about it? Not much beyond reporting.
One major vendor can produce a report that accumulates a company’s electric bills, air travel costs, vehicle costs and a few other carbon generating costs. That’s pretty much the extent of their solution.
What businesses need is a green solution that is smart enough to schedule production at facilities that can utilize hydro-electric or wind power instead of fossil-fuels. They need to know the right time of day to schedule production. Did you know for instance that electric power in some areas of the country is entirely hydro-electric at night while daytime power is a mix of fossil fuel and hydro-electric power? They need to know which machine tools are the most and least energy efficient. They still need financial accounting enhancements to better report the fully loaded cost of an item including carbon offsets and to recommend strategies to reduce these.
Businesses didn’t get these because ERP vendors would have to rewrite modules like Cost Accounting, Production Scheduling, etc. Even the algorithms used by the Production Scheduling module would have to be completely rethought. Sadly, making these changes is something that ERP vendors mostly chose to ignore. It was just too hard, too expensive and too disruptive to the old product line.
New ERP needs new metrics
The operational and financial language of the modern, digital firm is unlike that of the Industrial Age.
Have you’ve listened in on an earnings call for a subscription economy firm? They don’t talk about any of the metrics/measures I learned in college finance and accounting. They’ve developed a new lingua franca for their world. They speak of churn, monthly recurring revenue, customer acquisition costs, etc. The pure-play digital firms like Google and Facebook use new metrics as well. It is almost as though the financial report with which so many of us are familiar has been relegated to an afterthought.
Digital firms speak to how well they are creating network effect around their websites, offerings, etc. They monitor web traffic, conversion of web visitors into customers, increases in lifetime customer value, and much more. They’re not measuring how much of machining scrap is over/under plan this month or what their labor variance by part is.
New metrics are being created for the new economy and businesses can’t wait for accountants, ERP vendors or others to define them. Business and the economy march on even if some laggards choose to stay behind.
Another challenge is that the fuel behind digital firms has no consistent way to be valued. Accountants are befuddled as to how to value data yet data and other intangibles represent a massive amount of the value of a modern firm. Too bad it doesn’t appear in audited financial statements.
We need an ERP for data. In the second part of this story, I'm going to look at what this means and who is doing the work to drag us into the 21st century.
[i] “The next Big Challenge for CEO”, ChiefExecutive.Net, J.P. Donlon, September/October 2015
Image credit - via Acumatica (partner at time of writing)