Why We Were Excited About the Temp Nurse Staffing Market — but Ultimately Decided Not to Pursue It

“Sweat, Blood and Tears… and Market Validation!”

Temporary nurse staffing at hospitals happens in broadly two forms: per diem (by shift / day) and travel (longer assignments; usually over multiple weeks, and could be out of the clinician’s hometown/home state). More than a year prior to diving into the healthcare staffing space, my team had won the Cedars-Sinai Techstars hackathon for our on-demand per diem nurse staffing idea/prototype (2nd place by the judges, and 1st place in the audience vote!). Fast forward to Summer 2017, when me and my cofounder got busy researching the market, got initially excited about entering it, but ultimately chose otherwise.

From a high-level perspective, the market stats looked great. The per diem staffing market size was $3.0bn in 2016, growing at 5% y/y. Ancillary markets which we could tag on to our platform in the near-term were also fairly large: travel nursing $2.2bn (growing 7%), allied health $3.2bn (6%) and locum tenens $2.7bn (8%).

The market dynamics also looked very favorable for disruption — our understanding of the predominant market set-up: the market is highly fragmented with no nurse registry holding more than 11% market share — the largest players include AMN Healthcare (11% share), CHG (8%), and Cross Country Healthcare (6%). Many registries take commissions as high as 50%, effectively overcharging hospitals and underpaying their nurses, and do most of the matching process manually (phone calls, emails and some using faxes). Many smaller nurse registries were started by retired nurses who had a sizable personal network of nurses, and who also knew the staffers at local hospitals. There is some working capital required to start (and scale) a registry, and of course you’d need the initial network and connections, but the overall barriers to entry are fairly low — hence the scores of players in the market.

Our initial vision was to be an on-demand per diem staffing platform — and effectively replace nurse registries. The head nurse at a hospital would tap her phone screen and a qualified nurse comes over for the required shift (just like ordering an Uber!). For providers, it would mean no intermediaries, no managing / reaching out to multiple registries, lower rates, and intelligent matching (want a nurse that speaks Spanish, or knows a specific EMR? You can add it to the filters!). For nurses: no more waiting by the phone at 5am for the registry to call you for an available shift, no signing up for multiple registries and diligently waiting for their emails, higher pay, easier calendar updates for availability, and better matches with your preferences taken into account.

Our initial market research and validation efforts involved our own research, talking to other registries, nurses, hospitals and seeking general feedback from our network and advisors. The market validation and deep dive led to a lot of crucial discoveries that weren’t as evident from high-level research. We learned a great deal about MSPs and VMSs (through research, and by talking to registries and nurses) who acted as gatekeepers for the majority of the provider market. The provider sales cycles were annual — the provider / MSP would annually evaluate their list of approved registries and determine pricing for pay rates. We got to dig into the economics of running a traditional registry business and found out that high commissions did not always translate into high margins. Registries dealing with outdated VMSs were also drowning in data entry tasks (due to minimal interoperability/automation), in addition to their credentialing duties.

Armed with the initial research and feedback, we went back to the drawing board, and started “digesting” the information — evaluating the good, the bad and the ugly, and brainstorming how to play in this ecosystem, make our product stickier, and more valuable and differentiated than traditional agencies. This resulted in expanding our vision to include a SaaS platform for providers for predicting staffing needs and optimizing staffing levels, which would be tied into our staffing marketplace.

The long-term vision was that the platform could predict future staffing shortages and mobilize the internal floating pool or procure external talent accordingly. There was some precedent for this — a large hospital system was testing an internal staffing prediction model, and some healthcare analytics startups were [not-so-prominently] offering pilots of staffing optimization software.

We subsequently launched two initiatives in parallel — we started working on preliminary predictive models, and developing the proof of concept (PoC) for the staffing platform, while continuing to push for market validation for our latest vision. After multiple calls and meetings with head nurses, staffing groups within hospitals (large systems and smaller hospitals nationwide), others working on / researching / writing about hospital staffing optimization, agency owners, and some angel investors, we learned the following:

• There was no way to bypass the MSPs initially — and we couldn’t displace them as we just wouldn’t have the credibility to sign SLAs with the providers.
• This meant we start as an agency (a “2.0” version but a registry nevertheless), and work up our way into the CIO’s office to pitch the SaaS platform, and also eventually elevate to the MSP level on the agency side.
• Many hospitals didn’t have rich (or any) meaningful historical staffing data, pushing our timeline further back on implementing models for them.
• Which brings us to: most hospitals we could talk to did not care much about staffing predictions and optimization. They had much bigger fish to fry, and the data collection processes (for the predictive models) which would be added to their staff’s workflow was a no-go. Some large multi-facility systems had efficient enough internal pools (and usually a dedicated staff member to manage the pool), and did not consider the temp staffing process (or the high rates) a major pain point.
• In addition to not having the provider side locked down, our value proposition for the participants on the other end of the marketplace — the nurses and other licensed clinicians — had not evolved, and the current value proposition was not enough to command platform loyalty, hence, we’d still need to market to and acquire nurses on a per-shift / per-assignment basis. We envisioned this would translate into ugly marketing wars with other agencies (think: the Uber/Lyft fight) for nursing talent down the road, eroding our margins and ultimately voiding our “Agency 2.0” business case.

Hence, after weeks of market validation and PoC work, we ended up at crossroads and had to decide what path to take. Some of the options we saw in front of us:

1. Continue looking for hospitals who are passionate about solving this problem (we had reached out to hundreds of hospitals at this point) and in parallel, spend time and resources to build an MVP for the staffing predictions and optimization platform + simultaneously build out the staffing marketplace — we weren’t too excited about this path as we were looking for passionate early adopters who would be excited about testing the MVP out and potentially help us define the direction of our products/vision — we hadn’t yet found any from the multiple calls and meetings we conducted.
2. Build a standalone staffing marketplace and scale it to a critical mass, eventually gunning for the MSP role, and replacing MSPs, VMSs, and agencies — this was also a no-go for the most part given the challenges mentioned above — no loyalty and no real differentiation in the eyes of the providers and clinicians; and scaling would be challenging with the unit economics not so great. With this path, we also struggled with some fundamental questions — what is our long-term goal? Are we really “disrupting” the market / 10x better than the current solutions / other emerging digital players? Do we want to go big or build something to sell to a larger agency? At one point we even considered reaching out to PE funds to spearhead a “roll-up” of smaller staffing agencies — but ultimately it wasn’t what we had set out to do (which is build a meaningful and disruptive business)!
3. Move on — go back to providers and focus on other pain points or zoom into a specific pain point we may have already come across — and try to validate that. We had been looking at some other issues in parallel (mostly operational inefficiencies, back-office automation for providers and the like) which we could start evaluating. It ultimately came down to resource allocation — did we want to spend our time and money continuing to pursue the staffing market or look for alternate opportunities to create a disruptive business? As you may have guessed, we ended up going with this option.

Lessons learned: Although you will have most likely heard all or most of the below, for us — as first time founders — these were eye-opening lessons that we had only read about in books, articles and blog posts, but were now experiencing firsthand.

1. Is your problem a major problem for your customers, i.e. in the top 3? In our case, we concluded that having a better/cheaper/faster temp nurse staffing solution, and a staffing predictions & optimization platform was not a top 3 priority for most providers.
2. Can you reach your customers directly and if not, who are the gatekeepers? Can you become a gatekeeper? While there is sufficient struggle to get through to customers in most industries, the healthcare industry is particularly notorious when it comes to gatekeepers, and its painful for startups to get through to providers, payors and other stakeholders. In this specific instance, we had to deal with third party MSPs and VMSs, in addition to pitching to the CIO’s office and getting a buy-in from the HR/staffing/nursing groups.
3. Unit economics need to work favorably from early on, and you also need to consider working capital in your business model as it can become significant when scaling (for temp nurse staffing, we needed to pay nurses right after their shift, while many hospitals settled agency accounts quarterly). If you are modeling in economies of scale over the long term, you need to be clear on how/when/if the unit economics of the business would improve over time.
4. Customer loyalty is an important consideration, and lack of loyalty could mean material S&M expenses and CAC down the line. We internally realized that we preferred business models with longer-term customer commitments / lock-ins, and opportunities to build on top of those commitments (“land and expand” strategy). While there very well could have been an opportunity to creatively lock-in either or both sides of this market (and make our platform the default choice for nurses or providers to use in the long term), we couldn’t confidently point to one or more set of ideas that could lead to meaningful customer loyalty for our platform.
5. Market dynamics: One of the reasons we decided to not pursue the market was because we would be entering a crowded [maturing] market vs. an emerging market. Our digital staffing solution was not radically different than what many players had started doing, and what incumbent players could easily build (and the providers weren’t interested in our proposed radical solution!), and with low barriers to entry, the ecosystem will end up in a race to the bottom.
6. Importance of IP and building defensibility: While this point was not entirely new for us, the experience did reinforce the importance of defining and planning for defensibility. During my prior role at a hedge fund, we did meticulous due diligence on every potential investment’s defensibility — something that had been engrained into me from my investment days. We assumed our moat would have been built over time with a 2-sided marketplace, and took our eye off the ball on defensibility while in the weeds of our concept and product development (we also got feedback from some advisors and angels during our calls on the lack of meaningful IP in our vision).
7. Prioritizing tasks: I got to learn many new things during the process (building wireframes, deep dives into prediction models, creating explainer videos and other marketing materials, creating a bot to walk nurses through becoming independent contractors!), but I believe that we wasted a lot of time on many activities that did not directly result in market validation, which could have been avoided altogether.

Did we make the right decision? Only time will tell! I continue to come across new startups that have jumped into this market after we decided to pull out, but I am confident of our analysis, glad we made an informed decision, and excited about the other opportunities that opened up (and are opening up) since then.

Sources:

• Harris Williams Healthcare Staffing Market Overview 2016
• Staffing Industry Analysts