Trifurcating today’s HR software market
- Today's HR market is extraordinarily complex. Here is your guide to navigating the morass and how to avoid common pitfalls.
The Human Resources (HR) software market is a complex space. Three big piles of solutions/vendors have emerged and they each represent a different level of value potential and technical sophistication.
If you’re undertaking a software selection, clarity around the value and innovation you desire will simplify the selection effort.
Three types of solution
On one level, we see a number of maturing vendors I characterize as follows. They are...
- undergoing growth pains
- driving toward a liquidity event (e.g., IPO)
- appeasing board members and other investors who are demanding better fiscal management (a la Rule of 40)
- fighting pressure from activist shareholders
- rolling out cute, entertaining NLP (natural language processing) or analytic demos that won’t go into production for years
- auditing and litigating their install base as a cynical way to drive up LTCV (lifetime customer value)
- slowly introducing other innovations into moribund product lines
These vendors look like an old-fogy that’s still trying to look young and hip (think of your grandfather wearing skinny jeans and hipster clothing). The defining feature of these solutions is that most function like HR applications have always functioned except now the software runs on a newer platform and is accessible via smartphones. Think: version 2 of the status quo (Traditional Plus).
There’s a second group of vendors who aren’t tied to the past (Modern). These new firms are:
- creating solutions that were never tied to old, obsolete HR process and data models
- using big, EXTERNAL data to make their solutions quite valuable
- imaginative and solve problems for constituents other than the HR employee
The defining characteristic of these vendors/solutions is that they represent radically different ways to approach an HR process. Instead of ‘recruiting’, they may offer ‘recruitment marketing’. Instead of asking operational leaders to manually complete a candidate evaluation form, the software uses 25,000 visual cues found in a 10-15 minute video interview.
And finally, there’s a third group of vendors and innovations that are focused on faddish, interesting but unproven solutions (Fad/Hyperbolic). These firms are trying to solve ridiculously complex, multi-faceted problems (e.g., engagement, retention, etc.) with tools that are often opaque and lack falsifiable scientific proof.
The defining characteristic of these tools is that they hit on a hot issue in HR today but the science behind the solution is often thin or tangential.
HR software buyers need to shop wisely.
If software buyers choose Traditional Plus applications, they’ll likely find large HR/HCM suites with old-school, traditional (and out of date) process designs.
Yes, the new solutions might now be multi-tenant, run on public clouds and even support in-memory technology. But, while the new platform has the potential of doing great things with big data, dark data, unstructured data, etc., the core applications and the data models they utilize were not designed for this world.
The powerful analytics, algorithms and robotic process automation that could be there, generally, isn’t. There’s potential, but for the foreseeable future, the software will look and behave a lot like your current old-school software.
These solution providers are afraid of doing anything remotely different as a big change delta might scare their installed base. The hallmark trait of these vendors is to slowly, ploddingly move the customers they already have to a new future.
They move in very tiny baby steps so as to not startle their existing customers. You might get innovation from these firms but it might easily take a decade or more to manifest itself.
The Modern vendors offer solutions that are radical departures from past practices and processes. These firms imagine new data sources (How about those outside of the core HRIS?) and new ways to do accomplish a task. The re-imagined HR process is the most attractive part of these solutions.
Modern vendors move faster because they’re more concerned about the customers to come and not the ones they already have. They feel that they can take risks and they appeal to prospective customers who are willing to embrace some of that risk, too.
You will likely find these vendors don't offer a large HCM suite as most are too new to have developed that many applications yet. And, in all likelihood, you’ll have to integrate some of these modern solutions to at least one piece of an old-school suite. It will feel like someone’s trying to retrofit solar energy panels on your 1920’s craftsman-style home with its original knob-and-tube wiring.
It won’t be elegant but it could create a competitive advantage for your organization if the modern solutions are solving material issues for your workforce.
Big changes in recruiting technology have occurred in recent years. Look at HireVue (for video interviewing and their smart assessment technology), Smashfly (for recruitment marketing), Entelo (for access to a billion+ database of workers and technology that senses the best time to approach a potential candidate), or Pymetrics (for better matching job holders/seekers to the right role/position). Some of the innovators in this space are using big data at the beginning of the recruiting process. Some are using powerful algorithms and machine learning to predict potential candidate career mobility. If your firm still thinks of recruiting in terms of classified ads, headhunters and job boards, think again.
Some learning solutions are going through a rebirth and reinvention of late. Cornerstone’s curation of training content and its ability to determine the most appropriate NEXT course one should take are great steps forward.
The nature of training has changed such that workers want training on demand. Workers want the content indexed and time stamped so they can go directly to the exact point in the training where they need help. But having a learning tool that knows what skills you already have, what role you have in the firm, what your likely career path will be, etc. allows learning tools to be smarter if a vendor applies machine learning to the tool. You can expect more modernization and innovation to occur in the learning segment.
Performance appraisals are a necessary evil of HR and operations. While companies are still debating whether they should continue to conduct these, software vendors have experimented with scores of new ways to capture some of the valuable content that could go into a performance review. Software that captures snippets of feedback found in collaboration tools, emails, social media and other sources can be used in some of the newer performance management tools.
Modern solutions are often capable of serving non-HR personnel. These products are designed for operational folk within your firm as well as non-employees. The new users could include contractors, staffing firms, regulators, shareholders, students, and, people who weren’t even looking for a job. The products that are designed for non-HR people are fascinating because they make HR people re-evaluate why their online job application workflow requires applicants go through 14 screens of data collection plus spend hours correcting resumé parsing errors.
Many of the modern solutions possess:
- unconstrained, scalable technology
- massive, new data sets
- new process designs
And, these modern solutions require:
- different HR Skills (e.g., math/statistics skills, machine learning and social science knowledge)
- a culture of risk-taking and experimentation within HR
- an HR housecleaning to eliminate some of the diversity in core HR solutions in use at the firm. It’s too expensive to develop and maintain large numbers of integrations to a number of redundant applications at every plant, business unit, etc.
- a willingness to acquire and use big data
Somewhere, right now, an HR executive is speaking about how they need more engaged employees, how a wellness program will improve the company’s culture, why analytics are going to save the company, etc. While their heart is in the right place, many of the cure-all solutions out there will not conclusively deliver what these executives want.
At countless vendor demonstrations, we hear of solutions that purportedly identify the best people to hire, put on teams, identify potential career longevity, stop or slow attrition and even lower diabetes (no lie!). These touted benefits are usually unsubstantiated. Listeners are supposed to believe that if they only acquire this technology, all their culture, retention, engagement and other challenges will magically disappear. But there are some issues to overcome before these cure-alls can be deemed legitimate.
- The ‘connection’ between a research report and a technology solution does not always deliver value. Lots of vendors will reference a paper they’ve read in the Harvard Business Review or another reputable source. The gist of that piece might conclude that one company had better market uptake versus a competitor and that the winning firm had a great culture. This is problematic on two levels: it represents correlation not causality and it involves a sample size of only two companies.
- It’s extremely rare to find a study of any technology solution’s effectiveness and/or efficacy. Where are the double-blind studies with control groups to see if these new solutions really work? How do we know if we’re not seeing a placebo effect that causes firms to report benefits that really aren’t there or won’t last? The scientific proof for these solutions is frequently non-existent. While I ‘know’ toxic cultures should hurt the bottom line, I can’t ‘know’ if a specific vendor’s culture tool is actually working.
- The data many of these tools use is the same old HR data your firm has had access to for decades with mabe one or two new additions. Human beings are complex entities whose ultimate behavior and job performance are influenced by numerous non-HR system and/or non-work actions and data. A worker going through a hostile divorce, for example, may not have the best attitude some days and appear less than fully engaged at work. How is an engagement tool going to know this, especially if it utilizes months/years old HRIS data? Old payroll data is not going to tell you much about a person’s level of engagement.
- The ‘cuter’ the tool, the more likely employees will game it. There are tools that ask employees to indicate how they’re feeling right now and somehow that gets balled up with other data to determine someone’s level of engagement. A relative of mine always responds very positively no matter how he really feels as he doesn’t see any upside to telling the truth. It’s foolish to think that workers can’t see through these ‘solutions’.
I remember one day firing someone for cause. After this particularly contentious encounter, I went to give blood. The blood drive people refused to accept me as my blood pressure was too high following this HR event. Now, if a smartphone app had asked me, at that moment, how I felt, I would have entered a pretty dark response. It didn’t mean I was disengaged with the firm or some other negative/diabolical issue. It's just the case that at a given point in time, which could be fleeting or long-term, I wasn’t feeling great. If your tool always catches someone before their morning coffee, are they really disengaged or just under-caffeinated? Tools like this are like hand grenades – they might be generally on target some of the time but they can also wreak havoc on innocent bystanders.
- Engagement and culture can change a lot faster and more frequently than these tools. Just watch someone’s engagement when a close friend and colleague gets fired or quits. Many of the new tools have so few indicators in their algorithms that they will likely miss many potential issues, incorrectly assess others and never get better unless they capture new sources of information.
- Opaque tools are inexcusable. If a vendor won’t let you see inside the black box, how can you defend its recommendations? How can you correct and improve an algorithm if you can’t get into it? Walk away from any non-transparent tool. It may be too generic to do any good and/or its logic could trigger a hard to defend lawsuit. Caveat emptor.
- Many of the tools are backward looking. How many of us remember those mutual fund ads on television where the voice-over announcer reminded us that ‘past performance is no guarantee of future returns’? Machine learning, by definition, must find a prior pattern and then it will look for it to occur again. It can’t, for example, detect the unique set of circumstances or traits that could cause a major personnel problem that one anomalous worker could create. No matter what the salesperson says, these tools don’t ‘predict’ future events, they can only highlight when another similar set of data points has come together.
- Is partial success acceptable? A tool that detected one worker’s upcoming departure could miss dozens more. How good does something have to be for your firm to risk adoption?
There are some FAD solutions that could go mainstream. Those solutions have one key characteristic: they add measurable value. RPA (robotic process automation), NLP and chatbots technologies could be winners as these products find uses for a number of time-consuming tasks to which HR professionals must respond. If a tool can be programmed to understand an employee’s question (e.g., “How do I get started with benefits enrollment?”), then it can serve up the appropriate information back. Repetitive tasks and questions are prime candidates for this kind of technology.
HR departments will need to acquire new skills in machine learning, statistics, integration, big data analysis, social sciences, marketing, and many more disciplines. For HR to remain relevant, it has to have team members with skills that complement the new technologies. If your firm is reluctant to add headcount and budget to HR, then HR must adopt technologies that take away repetitive work and free up HR team members’ time to work on more strategic tasks.
Analytics is another area that warrants examination. Some analytic tools could become very successful if they demonstrably, easily, effectively and transparently solve a real business problem. To create a successful solution, vendors must:
- look at data beyond that which rests exclusively within the core HR solution
- make the underlying algorithms transparent
- be able to defend the logic in court and to regulatory bodies
- make these easy to integrate as no two customers have identical software portfolios
- create global analytic apps (This is no small feat as items such as culture or engagement can exhibit significant differences within a firm due to geographic, leadership, nationalistic, religious and other factors
The very best analytic solutions will likely be those that:
- don’t just report that something anomalous has happened
- predict when a non-standard event is about to occur
- anticipate the next 2-3 actions or information needs of a responder
- permit the responder to select an action and have the workflow built-in to complete the task
So, with some tweaking, some of these faddish solutions can evolve into something worthwhile.
I often counsel clients to buy/subscribe to software that is technically current today and has a great runway of future innovations ahead of it.
In today’s HR market, there are products that are only now becoming technically current and could easily a decade away from delivering anything novel. There are also very modern products from smaller firms. Which of these companies will morph into a giant killer remains to be seen. And then there are firms that trigger animated discussions but will likely hit resistance from the executive committee if a compelling and believable business case can’t be made.
Buyers should be happy to see the breadth of solutions and level of innovation in the HR space today. This wealth of choices and approaches points to a vibrant market. HR buyers will need discipline and a discerning eye to choose wisely. Few firms can afford to try everything out there so a framework for selection solutions is needed.
The three buckets in the preceding paragraphs generally apply to specific solutions but might apply to the entire company. Also, note that a solution can start in one bucket and, in time, move to another. The software industry is rarely static so make your selection assessments based on current product knowledge.
This is an unclear space and one that warrants care.