I have been looking forward to Infor's user conference, that has started today in Orlando, for a while. The company has been talking and hinting at their microvertical strategy a lot - but never in a public forum in combination with their new products. So today was going to be interesting, and it was:
[If not familiar with Infor - read today's well timed USA Today piece about Charles Philips here.]
![]()
To differentiate themselves, Infor has chosen not just a vertical strategy - but a microvertical strategy. Go beyond the 20+ or so industries all the way to 200+ microverticals. That's a very clever differentiation, driving business automation further than SAP and Oracle have done. Infor CEO Charles Philips drove this argument home with a slide of labeling enterprise application eras:
And I very much like the slide, as it clearly outs the design mantra of current enterprise applications: There is only one monolithic way to automate the business. To allow different functionality, the vendor may allow configuration of the monolithic application. If the scope of the configuration does not go far enough though, you need to customize (or with SAP modify) and then the ugly TCO monster rears it head.
And Infor has done good work to understand the underpinning of such a strategy - with its Ion middleware. Not enough details on Ion were presented in the keynote to assess it [hope to be able to do that sometime soon] but it looks promising.
I give Infor high marks for sticking with standards based on OAGIS / XML, which should make it easier to operate with other applications from 3rd party vendors. And its good for the #3 and beyond vendors to be able to hide behind standards.
But not only do you need to make data flow and make APIs available (still a lot of questions how that works) - users need to be able to co-operate across their current application boundaries, which requires a interaction / social / collaboration platform, which Infor has unveiled earlier already with ming.le.
ming.le serves practically as an interactive communication tool, as the presenters made very clear, that key identifiers such as the customer ID, product ID etc are passed along with the ming.le interaction.
But then users may find themselves in enterprise applications, that they are not familiar with, so it's key to unify the user interaction across the many Infor products, and that job falls to project Soho, an HTML5 based set of UI tools, forms and standards, that achieves - at least at screenshot and demo level - a very high conformity of usability across the products.
Last but not least - and that was brand new news today - Infor needed to address the reporting / BI challenge. With separate silos of information that are glued together with XML messages, and despite a social collaboration platform and harmonized UI across them, enterprises still want to report on cross silo data. As none of the silos (or products) have all the data, Infor had to come up with a solution and did so with a very elegant approach:
The design is for Ion to puts copies of all relevant data - probably not only what it carries through for transactions across the different products, but also beyond that - into one large storage in the cloud. Interestingly Infor choose Amazon's Redshift (former ParAccel) as the technology to power the new SkyVault solution. A smart move as it gives Infor flexibility and avoids significant investment in capital for their own cloud - given the uptake of SkyVault with customers is at least a bit uncertain in the next quarters. Needless to say the low cost of Amazon's Redshift offer is of appeal to Infor, as Infor does not want to loose too much share of wallet with their customers with this architectural decisions. $s for Infor, cents for Amazon you may think.
So amidst all this enabling technlogy, Infor also upgraded and revamped the underlying applications from Infor 10 to Infor 10x. To make it into the 10x club, applications had to fulfill a number of conditions, namely to leverage the above investments, as the below slide sums up well:
Moreover Infor announced that a number of their applications will be available through IBM's Smartcloud services. A good move by Infor, as it again alleviates heavy capex investment (that should go to product development in this phase) and offers a large brand name to create trust into the cloud offering, that some of the 70000 Infor customers may not yet have. The other compelling feature of choosing IBM's cloud products is, that they fully support private, public and hybrid cloud as there is no way for Infor to foresee how their customers will migrate their applications to the cloud.
And finally Infor had an interesting investment around a (fittingly) (micro)vertical cloud for equipment vendors. Taking the M3 product, leveraging an investment that NTT Data has done together with Velocity Technology Solutions on IBM cloud services, there is a compelling cloud offering for companies that rely heavily on equipment. Fittingly the offer has been endorses by Caterpillar dealers around the world.
Unrelated to the architecture announcements, it was good to see the strong commitment to design thinking, with creating an in house design team with Hook&Loop. The impact is already seen in the new product releases and is encouraging. I was impressed by the example of the re-designed hotel check-in application. We will see over time how more continuous improvement can be instilled in the Infor applications. But it's a very promising start.
Pretty much all the Infor products have their inception as a product, their basic design, data structure and capability to automate industry best practices from over 10+ years ago. That by itself doesn't have to be bad - but creates significant technical debt on Infor and Infor customers. The good news is, that some of them, like the customer reference proudly talking about their AS/400 solution - don't even mind. The bad news is, that you cannot build a next generation, best practices enterprise system.
Infor - unwillingly of course - even documented that in the two key note demos. Of course ming.le was used to tie different Infor acquired apps together- but what was missed, was that a business automation system in the 21st century, should have automated flows a little differently and better:
![]()
[Update April 23rd 23:58 PST:] Don't miss my takeaways of the Day 2 keynote here.
[Update April 25th 8:03 PST] - You can watch the replay of the keynotes here.
[If not familiar with Infor - read today's well timed USA Today piece about Charles Philips here.]
The Good
If you follow enterprise software as a customer, vendor, partner, analyst or blogger - you have to pull for Infor. The SAP and Oracle duopoly really needs a run for the money - for the better of the market and the industry. And Infor is doing a good job, labeling themselves as the greatest startup around, given the recent over 1 Billion investment, the impressive hiring of talent and the ambitious re-architecting of platform and applications. |
| Inforum 2013 Keynote - Day 1 |
To differentiate themselves, Infor has chosen not just a vertical strategy - but a microvertical strategy. Go beyond the 20+ or so industries all the way to 200+ microverticals. That's a very clever differentiation, driving business automation further than SAP and Oracle have done. Infor CEO Charles Philips drove this argument home with a slide of labeling enterprise application eras:
 |
| Inforum 2013 Keynote - Day 1 |
And Infor has done good work to understand the underpinning of such a strategy - with its Ion middleware. Not enough details on Ion were presented in the keynote to assess it [hope to be able to do that sometime soon] but it looks promising.
 |
| Inforum 2013 Keynote - Day 1 |
But not only do you need to make data flow and make APIs available (still a lot of questions how that works) - users need to be able to co-operate across their current application boundaries, which requires a interaction / social / collaboration platform, which Infor has unveiled earlier already with ming.le.
 |
| Inforum 2013 Keynote - Day 1 |
ming.le serves practically as an interactive communication tool, as the presenters made very clear, that key identifiers such as the customer ID, product ID etc are passed along with the ming.le interaction.
But then users may find themselves in enterprise applications, that they are not familiar with, so it's key to unify the user interaction across the many Infor products, and that job falls to project Soho, an HTML5 based set of UI tools, forms and standards, that achieves - at least at screenshot and demo level - a very high conformity of usability across the products.
 |
| Order Screen after applying Soho - Inforum 2013 Keynote - Day 1 |
 |
| SkyVault - Inforum 2013 Keynote - Day 1 |
So amidst all this enabling technlogy, Infor also upgraded and revamped the underlying applications from Infor 10 to Infor 10x. To make it into the 10x club, applications had to fulfill a number of conditions, namely to leverage the above investments, as the below slide sums up well:
 |
| SkyVault - Inforum 2013 Keynote - Day 1 |
Moreover Infor announced that a number of their applications will be available through IBM's Smartcloud services. A good move by Infor, as it again alleviates heavy capex investment (that should go to product development in this phase) and offers a large brand name to create trust into the cloud offering, that some of the 70000 Infor customers may not yet have. The other compelling feature of choosing IBM's cloud products is, that they fully support private, public and hybrid cloud as there is no way for Infor to foresee how their customers will migrate their applications to the cloud.
And finally Infor had an interesting investment around a (fittingly) (micro)vertical cloud for equipment vendors. Taking the M3 product, leveraging an investment that NTT Data has done together with Velocity Technology Solutions on IBM cloud services, there is a compelling cloud offering for companies that rely heavily on equipment. Fittingly the offer has been endorses by Caterpillar dealers around the world.
Unrelated to the architecture announcements, it was good to see the strong commitment to design thinking, with creating an in house design team with Hook&Loop. The impact is already seen in the new product releases and is encouraging. I was impressed by the example of the re-designed hotel check-in application. We will see over time how more continuous improvement can be instilled in the Infor applications. But it's a very promising start.
The Bad
Even if you invest around 1B in your enterprise company, and can show a significant shift forward, with the over-quoted massive investment in products (over 850 new developers) - you are still tying together old products - with modern tools.Pretty much all the Infor products have their inception as a product, their basic design, data structure and capability to automate industry best practices from over 10+ years ago. That by itself doesn't have to be bad - but creates significant technical debt on Infor and Infor customers. The good news is, that some of them, like the customer reference proudly talking about their AS/400 solution - don't even mind. The bad news is, that you cannot build a next generation, best practices enterprise system.
Infor - unwillingly of course - even documented that in the two key note demos. Of course ming.le was used to tie different Infor acquired apps together- but what was missed, was that a business automation system in the 21st century, should have automated flows a little differently and better:
- The 1st example was the classic front office to back office demo: Sales rep realizes that product will be late for a key customer, he contacts the back office (granted, nicely done 21st century style a ming.le @production message), production manager finds out why, it's a planned maintenance, moves the planned maintenance and all users are heroes, the customer wins.
But: The plant maintenance should not have been scheduled in the 1st place - as it disrupts the ETA of a key customer's products. All what we have seen was a lot of messages and rescheduling (I may say dust) about something a 21st century business app should not have let happen in the 1st place).
[Ironically this will be a sales argument of Oracle and SAP against Infor, given they operate one one information set / schema - and Infor does not. Why show this demo at Inforum? No idea.]
- The 2nd example was another classic I have seen in many incarnations: The urgent search for a spare part, again powered by ming.le, the obligatory internal check if other locations have the spare part inhouse and you could transfer them internally, and then the purchasing manager buys from the 2nd highest bidder, because they can deliver the critical spare part today - not later. Then we see how the VP of purchasing gets a message that the lowest bidder standard process was violated, but after back and forward with the purchasing manager, approves the purchase. Heroes all over again.
But: A 21st century application should have allowed the purchasing manager to notify the VP why the lowest bid was not selected. Ironically it looks like the products could have done that, but the demo didn't. But even better the procurement rules should have allowed for exceptions to the lowers bidder rules, based on delivery speed, criticality, business impact etc .
Don't want to be too picky, but demos matter, as they show the capability of products, vision, direction etc. But it's intuitively clear that tying together past business practices with modern tools, does not allow for tomorrow's business processes.
My second concern is, that Infor needs to pick where to invest - and where not. I picked up a tweet, that ming.le runs on top of Microsoft Sharepoint. I have no further details and confirmation, but if that is the case, licensing may be so prohibitively costly, that it may never take off. Similarly using Amazon redshift is a smart move to save infrastructure capex, but longer term no enterprise software vendor can afford not to have significant IP in their BI offerings. You do not need to build something like Hana, but you need to have enough chops in the game to nail BI / Analytics vis a vis your transactional applications.
My last concern in the The Bad category is, that all the acquired Infor applications sit on very different, some on pretty old technology stacks. When the Infor executives transport their microvertical message through examples of deep vertical automation, so deep that it makes a monolithic vendor developer cringe, we cannot oversee that all that deep automation is being built on old platforms. And though Infor has done a very good job to access that functionality with their modern Ion platform - it's still an older technology platform with lots of old code.
The Ugly
Infor touted many very impressive numbers about the ongoing transformation:
 |
| Order Screen - Inforum 2013 Keynote - Day 1 |
But the scope is simply huge, I'd say breathtaking:
 |
| Order Screen - Inforum 2013 Keynote - Day 1 |
In the past history of enterprise automation many well starting verticalization efforts have fizzled, due to architecture issues and lack of development capacity. Remember the Oracle CPG offering? It was a lack of architecture and the move to thin clients on the underlying horizontal applications that required a reset of that effort. Or all the SAP verticals announced in the 2nd half of the 90ies that fizzled for lack of customer uptake to finance the sheer amount of manual work to keep them up to speed?
And while I think Infor has a good strategy and approach on the architecture side with Ion, I am very worried about the development capacity.
Yes Infor has 3500 developers. Let's assume the two large blue boxes above came for free and all capacity would be available for building business functionality. That means still 21 horizontal apps flavored by 5 verticals. Making it 105. Let's discount some for reuse and make it 70 specific microvertical / horizontal flavors which require investment. That would make it 50 developers per offering. Sounds like a lot for the start, but take away 20 heads for product management, quality assurance and documentation. Now factor in that Infor is not a pure SaaS vendor, and allows on premise installs, which means customers will be on older releases. Which means different code lines for error correction. So the 30 boil down to 10-15 for new code. Still nice to have - but in my experience short of moving the needle permanently. Now don't forget that the two blue boxed above are not free and will take a chunk of development capacity out of the above numbers.
Infor may say, that they will grow themselves out of the problem, growing revenues and hiring more developers, and I am pulling for them. But I think they will not be able to maintain this very wide scope. It will be a good sign in my view, when Infor will stop some less selling microverticals. That will be painful for the affected customers, but good news for the remaining customer base, as they should see higher per microvertical investment.
If you think this through, it means that a former Baan, Lawson, epiphany customer, who was worried about their vendor viability back then, is not in a significantly better position now. Granted the investment is impressive, the sharing of the technology across the products is more investment any of the vendors could have single handedly done before - but the problem - how many customers share this vision on how to run a business in a microvertical - remains the same.
MyPOV
Impressive work done by Charles Philips and team in about 18 months. Infor has done all the right things you can do in this short amount of time. If the underlying products are still in tune with 21st century business practices, if their technology allows still for productive coding of new business practices and if Infor will be able to grow the customer base to make investment challenges on a microvertical level disappear - remains to be seen.
A lot of Good, some Bad that can be corrected and some concerns in Ugly that need Infor needs to address.
In the meantime, I am pulling for Infor, as it will help the enterprise software market to give SAP and Oracle a run for the money.
And that really no one will show up with the a slide like below again:
[Update April 23rd 23:58 PST:] Don't miss my takeaways of the Day 2 keynote here.
[Update April 25th 8:03 PST] - You can watch the replay of the keynotes here.
P.S.S. And a good conversation of Dennis Howlett with Ray R Wang on today's takeaways: