Gartner: 96% of customers are getting real value from RPA? Really?


Last year we couldn’t help ourselves revealing our lovely Gartner analyst friends, via the voice of Chief of Research and Distinguished Analyst, Fran Karamouzis, declaring, “3 million of us will be supervised by robobosses by 2018“.  

So, while many of us are counting down the last few months enjoying our last experiences of having human bosses (or maybe some of us will actually prefer a robot), we can now breathe a huge sigh of relief that a whopping “96% of clients are getting real value from RPA” (Robotic Process Automation).  And not only that, RPA is thriving at a “satisfaction level greater than anything Fran has seen in her 17 years at Gartner”:

I personally would love to meet this incredible cross sample of delighted clients Fran has had the good fortune to interview, seeing as we’ve been covering the emergence of RPA for nearly 5 years and this space is still at a very early phase of (sometimes) painful RPA experimentation, as enterprises figure out how to scale these tools, govern them and learn how to integrate them with other applications using scarce technical skills, while dealing with very challenging change issues.  

At HfS, we just came off a very intense day with 60 enterprise clients tinkering with RPA, and can officially declare that 96% of them are definitely not in love with their experience.  In fact, only a handful are making real progress, while the majority lack a cohesive governance program to get this stuff working on even a few rudimentary processes.  At HfS, we estimate, from our extensive ongoing research, that about half of today’s RPA implementations are, so far, making some progress, while even Ernst and Young’s new RPA report declares it has seen 30-50% of initial RPA implementations fail. (And this McKinsey piece entitled “Burned by the bots: Why robotic automation is stumbling”  has since been published… well worth a read).   

Why claiming 96% of RPA customers are seeing real value is plainly ridiculous 

Several of the RPA solutions vendors are painting an over-glamorous picture of dramatic cost savings and ROI. RPA software firms are claiming – and demonstrating – some client cases where ~40% of cost (or more, in some cases) is being taken off the bottom line. While some of these cases are genuine, there are many RPA pilots and early-phase implementations in the industry that have been left stranded because clients just couldn’t figure out the ROI and how to implement this stuff. This isn’t simply a case of buying software and looping broken processes together to remove manual efforts… this requires real buy-in from IT and operations leaders to invest in the technical, organizational change management, and process transformation skills. 

Several RPA clients cannot scale their solutions and are aborting implementations.  One solution in particular, which featured high in many analyst scatterplots, has recently suffered the ignominy of not being able to scale at the level it needs, with several of its clients and projects being either aborted or moved onto other solutions

Buyers are backed into a corner with broken delusions of automation grandeur as their CoEs fail. Buyer leaderships are being fed all this rosy information and are under incredible pressure to devise and execute an RPA strategy, with some sort of set of metrics, that they can demonstrate to their operations leadership.  Many are quickly discovering they simply do not have the skills inhouse to set up automation centers of excellence and are frantically turning to third parties to help get them on the right track.

Outsourcing consultants are selling RPA before they can really deliver it. Sourcing advisors are claiming they are now “RPA experts” who can make this happen, while struggling to scale up talent bases that can understand the technology and deal with the considerable change management tensions within their clients.  RPA is murky and complex, and not something you can train 28-year-old MBAs to master overnight.  Meanwhile, we are seeing some advisors simply do some brokering of RPA software deals for small fees, only to make a hasty exit from the client as they do not have the expertise to roll-out effective implementation and change management programs. 

RPA specialist consultants few and far between. Pure-play RPA advisors are explaining this is not quite so easy and requires a lot more of a centralized, concise strategy.  There are simply not enough of these firms in the market, especially with Genfour having been snapped up recently by Accenture. With only a small handful of boutique specialists to go around, these firms can pick and choose their clients and command high rates.

Service providers are setting the pace, but will destroy each other in the process, making it challenging to source the right RPA capabilities. Service providers are claiming they can implement whatever RPA clients need, but are not willing to do it at the expense of reducing their current revenues. Meanwhile, smart service providers are aggressively implementing RPA into their own operations to drive down their delivery costs and reduce their own headcount, and many are already claiming 10-20% of their delivery headcount has been reduced. So we can expect to see providers aggressively attacking competitive clients with automation-led solutions that should create unbearable pricing pressures for service providers looking to retain the talent they need to implement this stuff. Hence, services providers will be hell bent on destroying each other and the winners will be those who eventually succeed in winning more work than they lose amidst all the destruction. 

Half of enterprise buyers want help from their service providers when it comes to RPA and cognitive.  When we privately polled 60 senior outsourcing buyers, at the recent HfS New York Summit, on what would improve the quality and outcomes of their current services relationships, the answer was pretty conclusive – half want to work with their providers to rollout their automation and cognitive roadmaps, begging the question why half of this famous 96% of RPA customers feel they need help from third parties if they are already so satisfied with their current RPA experiences?

The Bottom-line: It’s time to stop pandering to the hype merchants and get real about the true challenges of RPA

The biggest issue threatening the real progress of RPA is the sheer deluge of misinformation that is being churned out to the poor unsuspecting customers.  I cannot tell you how many have called us up at HfS bemoaning the fact their CFO has just returned from a conference and wants a piece of this 40% savings from recording manual processes in a digital loop.

Now, it may simply be that Fran has been lucky enough to have been spoon-fed references from only those highly successful clients of RPA, hence her unadulterated exuberance of its proven value. Sadly, I don’t get to wallow in such a haven of client success, and seem to have all the clients who are struggling to get this stuff moving call us up and come to our summits to share war stories.  However we conclude this latest little debacle in the unraveling of the RPA odyssey, it is clear is our leading analysts need to get their research right and owe our industry the real facts, not the puffed up hype to excite the marketeers and sales people in the software and services firms.  RPA impacts jobs, it impacts process effectiveness and can cost you dearly if you mess it up.  It needs to be handled with intelligence and diligence.

Posted in : Confusing Outsourcing Information, Robotic Process Automation



Leave a Reply

Your email address will not be published. Required fields are marked *

  1. Phil –

    Full respect for calling this out. The hype is endless and a 96% data point for RPA satisfaction pure science fiction.


  2. Thank you Phil for your solid research and informed perspectives. Living in the RPA trenches for the last five years I’ve seen it all (I still fondly remember the Robotistan paper – it’s what turned me on to RPA before it was called RPA). I marvel at how badly enterprises want to believe in a miraculous quick fix. And how quickly they will accept hyperbole, hype and and misinformation. RPA is a tool in the enterprise transformation toolkit. And, nothing is quick or easy about true enterprise transformation. But, done right (with the right partner, and the right software) enterprises can achieve real, meaningful value. It’s hard work. But what worth doing isn’t?

  3. I can’t agree more. Absolutely valid point placed and we should really advise the so called “analysts” to dig deep and share facts, not hypes and Acronyms.
    Coming to the topic, RPA promises lot – if done right! There is no short cut to success. At times I can’t stop but feel many organisations are considering it to be the silver bullet for all their problems. It may definitely serve to a great extent, but we can’t deny that it will require clear planning, execution and rollout. Not exactly recording a few steps and plying it over repeatedly. 🙂

  4. Phil, thanks for digging into the conversation on this. I don’t have enough data on the statistics but I agree that the organizational learning curve is not insignificant. It amplifies the need for a OneOffice approach with all the tensions that come along with that transformation.

    Outsourcing providers have a unique opportunity here to partner with their clients and help them through this journey. As you point out, customers are more than open to this approach. A call to action for those that are not there yet!

    @Ian, well put sir!

  5. Some parallels here to what happened 10y ago in transitioning Marketing function to become MarTech. When analytics and content mgmt came together in what is now known as customer experience space, marketers were also failing with customer experience – “personalization does not work”, “multi-site management does not work”, “a/b testing does not scale”, etc. This was precipitated by their then current partners (ad agencies) trying to do complex and thoughtful tech work and failing to scale or even failing to cover the whole lifecycle. Agencies later acquired tech firms to build their capabilities in execution and sustaining and space matured. Same seems to be happening in Operations now – operations advisory firms have the ear of ops folks and are trying to do thoughtful tech work and are failing. Just ask how many implementations Big 4 did that are actually in true sustaining phase. A lot of what is being built is not sustainable because it is done by firms who have done any tech-enabled work at scale before. Just like we saw SIs buying agencies and agencies buying SIs for martech we will see SIs buying ops consultants and ops consultants buying SIs in Automation.

  6. I’d put forth that those that did the due diligence up front and set matched expectations are receiving value and ROI. If the POC was entered with just hopes, they’re no doubt struggling

  7. Phil –

    I guess if 9 and 6 are grouped in some other way then it may be possible for example 9/6 or (9-6) 🙂

    Maybe the early adopters who have spent years customising the framework may have achieved certain value. I agree 96% appears to be quite high though!

  8. Ian – Very true. RPA is hugely powerful with great potential but only if it is rolled out as part of a greater strategy through a CoE approach with strong governance. It needs to be underpinned by solid process optimisation, waste elimination and standardisation and it certainly is not a plug and play solution but requires expertise, leadership and lots of hard work.

    Frank Smits

  9. I was surprised when I saw that 96%. I’m glad Phil Fersht had something to say about it.

  10. We are hearing less than 45% continuously – running our own poll – if response is big enough will publish

  11. Fully agree that RPA can deliver great results when implemented in the right way and in the right places but this is more difficult than it actually sounds.

  12. Marketing hype and reality don’t always equal. I like staying with the real facts when I do RPA demos. Doesn’t pay to fake it.

  13. Phil,

    Am so pleased you called out this out. Several of the early adopters are quickly realizing that some of these software tools are hard to scale, have security flaws and need much tighter governance than they initally were sold.

    I hope we can all get past the naive hype and focus on the real issues about how to develop an effective automation strategy and focus on more than fantasy cost reduction,


  14. Some of the challenges you pointed out are indeed valid and I see on a daily basis. However the RPA and Automation market is till very murky. We have a large number of successful customers that have build automation teams within the enterprise and are rolling out an ever increasing number of use cases across the organisation from ERP systems to digital apps.

    Through automation we are seeing not only cost savings but the generation of new ideas massively contributing to top line revenues.

    From my perspective, the definition of RPA should not be limited to tools that automate the front-end. To be successful companies need to embrace an automation mindset that may consist of organisational changes both structurally/politically and culturally, improve expertise on internal processes and invest in fit-for-purpose tools from front-end robotics, back end ITPA automation (also a type of RPA), Service Orchestration (self service) and even things like testing and release automation tools. Layer this with an enterprise wide, highly scalable, robust orchestration platform and you can realise significant ROI.

    I have seen a number of SI’s in the US that are taking this approach and have been able to provide their end customers with better and faster service while lowering their costs.

    With the prospect of AI (still a bit of a hype at the moment) we should see additional use cases being deployed throughout the enterprise. Perhaps policy based workload placement (i.e. on Prem or public cloud based on business policies) or auto analysis of support tickets to determine what service / remediation process needs to be fulfilled.

    Front-end GUI automation is limited and doesn’t solve real world organisational process challenges. I would challenge everyone to think bigger. Not just RPA but Automation as a platform throughout the organisation. Organisational change, education, process expertise and guts are all required to make it work. Stop thinking a magical tool will solve your problems – only you can solve your problems.

  15. Hi Phil,

    Thanks a lot for your blog post. Here at Contextor we agree with your point. There is no need to over-hype RPA, because it could cause huge disappointments later on.

    From what we have seen during the last decade, “Attended RPA” is generating more ROI on quick projects and far less delusion than many “Unattended RPA” projects…

    Our return of experience, after having installed about 100,000 RPA bots worldwide, 90% of our clients doing Attended RPA, is detailed on our blog :
    – “Robotic Process Automation: what is it about?”
    – “Robotic Process Automation: a sticking-plaster solution?”

  16. Mark,

    We at Cortex agree with you that Robotic Process Auotmation should be a much larger field, but unfortunately,.RPA has been successfully usurped by the Screen Scraping brigade and dumbed down to the fragile digital mimicry of button pushing… and it looks like it is never coming back.
    We prefer the term Intelligent Automation instead, as Tom Reuner and now Gartner(Grr) use. It helps define the conversation.. A Robot is only assumed to have some kind of decision making power, and often is merely an automaton doing dumb repetitive action. So we could either spend large amounts of marketing dollars to try and redefine RPA as Automaton Process Automation “APA”, or move on to define a new space; the smarter “Intelligent Automation” which must have intelligence and automation by definition.
    Our dollars are on Intelligent Automation as they have been since 2003!
    The evolution of automation is bring decision and action together, leaving automatons in our dust 🙂

  17. Fully agree with Phil Fersht: RPA isn´t easy and this kind of ridiculous statement isn´t helping Gartners reputation

  18. Thank you Phil for your solid research and informed perspectives. Living in the RPA trenches for the last five years I’ve seen it all (I still fondly remember the Robotistan paper – it’s what turned me on to RPA before it was called RPA). I marvel at how badly enterprises want to believe in a miraculous quick fix. And how quickly they will accept hyperbole, hype and and misinformation. RPA is a tool in the enterprise transformation toolkit. And, nothing is quick or easy about true enterprise transformation. But, done right (with the right partner, and the right software) enterprises can achieve real, meaningful value. It’s hard work. But what worth doing isn’t?

Continue Reading