Healthcare technology is usually sold one problem at a time. A practice needs a better way to communicate with patients, manage memberships, document visits, handle scheduling, process payments, or support clinical decisions, so a new product enters the picture with a very specific promise. Save time here. Reduce friction there. Automate this task. Make that workflow easier.
Individually, each promise can be completely reasonable. The problem is that physicians do not experience those products individually. They experience the stack.
That distinction became much more interesting to me after working on My DPC Story’s 2026 Battle of the Support Stack. We asked 131 practicing DPC physicians to rate the software they actually use across nine categories. We also asked a deliberately difficult question: if you could keep only one non-EHR tool in your practice, what would it be?
A total of 108 physicians answered. After consolidating obvious naming variations, Spruce received 21 votes, Hint received 20, and OpenEvidence came in third with 13. The ranking itself was useful, but the explanations were even more revealing. One physician described communication as the “backbone” of the practice. Another called Hint the “bones of the business.” Different products. Different jobs. Very similar language.
What those responses suggest is that the tools physicians are least willing to lose are not simply useful. They have become infrastructure. They are woven so deeply into how the practice operates that removing them would disrupt the business itself.
I think that distinction matters for healthtech companies trying to sell into independent practices, because a new tool does not enter an empty environment. It enters a practice that already has software, workflows, habits, integrations, contracts, support relationships, and existing points of friction.
Every new product may solve a problem, but it also introduces something else that has to be managed. Someone has to choose it, configure it, learn it, troubleshoot it, maintain it, renew it, and figure out what to do when it stops talking to something else.
In a large health system, some of that complexity can disappear into IT, procurement, operations, or implementation teams. In a small independent practice, it often cannot. The person evaluating the software may also be the person seeing patients that afternoon.
That changes the value equation. A tool can save time inside one workflow and still add complexity to the practice around it. The relevant question is not simply whether the product creates value. It is whether the product creates enough value to justify becoming another thing the practice has to manage.
That is also why the way we market healthtech can be misleading. Software is usually evaluated category by category: best patient communication platform, best membership software, best AI scribe, best scheduling tool, best decision-support product.
Those categories are useful for buyers and vendors, but they are not necessarily how the physician experiences the workday. A physician does not move neatly from the “patient communication category” to the “membership management category” to the “clinical decision support category.” A patient sends a message. A refill needs approval. A card fails. A lab comes back. A new patient needs to enroll. A note needs to be finished. A specialist record needs to be found.
The work crosses categories constantly, which means every additional product creates another boundary the practice has to manage. Another place where information may need to move. Another interface to remember. Another vendor relationship to maintain. That cumulative burden rarely appears in an ROI calculator, but it is still part of the product experience.
The obvious answer is integration, and good integrations absolutely matter. If two systems exchange information reliably, the practice does not have to manually bridge the gap between them. But integration is not the same thing as consolidation.
Two integrated vendors may still mean two contracts, two support teams, two renewal decisions, two interfaces, two product roadmaps, and two places where something can break.
That does not mean every healthtech company should try to become an all-in-one platform. Specialized products exist for good reasons, and a company that tries to do everything can easily become mediocre at everything.
But it does change how I think about product expansion. The question is not always, “What new feature can we add?” Sometimes the better question is, “What piece of the customer’s existing stack could this feature allow them to remove?” Those are very different product strategies.
Healthtech companies understandably spend a lot of time watching competitors in their own category. If another communication platform launches a new feature, the product team notices. If another AI scribe changes pricing, sales hears about it. If another membership platform improves onboarding, everyone pays attention.
But for a small independent practice, the real competitive set can be broader. Your product may be competing with the physician deciding, “I already have enough software.” That is a much harder competitor, because now the product does not simply have to prove that it is better than another tool. It has to overcome the operational cost of adding anything at all.
That is why the Support Stack findings made me think differently about the products physicians described as indispensable. The interesting question is not only why they were adopted. It is how deeply they became embedded in the practice.
Once a product becomes part of the basic operating infrastructure, adjacent functionality may have disproportionate value. Not because physicians necessarily want one company to do everything, but because getting one more capability from a system the practice already depends on can feel very different from adding another vendor to get it.
That changes the value proposition. “More features” is not especially compelling on its own. Neither is “all-in-one.” The more interesting promise is reduction: one fewer manual handoff, one fewer place to enter the same information, one fewer system staff need to learn, one fewer bill, one fewer integration to maintain, one fewer vendor to call when something stops working.
That is operational value too.
For companies selling into independent practices, I think it deserves more weight. These practices are not miniature hospital systems. Their constraints are different, their staffing is different, and their tolerance for implementation burden is different.
The best product is not always the one with the longest feature list. Sometimes it is the one that makes the rest of the stack simpler. And sometimes the most valuable feature a healthtech company can build is the one that lets its customer cancel something else.
Because your competitor is not always another product. Sometimes it is the cost of becoming one more thing the practice has to manage.
SOURCES & EVIDENCE
Primary evidence
My DPC Story. 2026 Battle of the Support Stack: 131 practicing DPC physicians across nine software categories; 108 answered the keep-one-non-EHR-tool question. After consolidating obvious naming variations: Spruce 21, Hint 20, OpenEvidence 13.
Evidence boundary
The Battle of the Support Stack directly supports the infrastructure/indispensability finding. The broader vendor-burden, consolidation, competitive-set, and product-strategy argument is the author’s interpretation; the survey did not directly ask physicians whether they wanted fewer vendors.
EDITORIAL STATUS + LINKEDIN PUBLISHING NOTES
Article 01 • Scheduled Sep 10
Your Patient May Have Already Had a Consultation Before They Meet You
LinkedIn article commentary
The patient walking into an exam room may not be starting the conversation from scratch anymore.
They may have already spent an hour asking AI what their symptoms mean, learning the vocabulary, testing possibilities, and figuring out what they want to ask.
I dug into new research on what happens when that conversation follows patients into actual medical care.
Article 04 • Scheduled Oct 14
Your 4.9-Star Rating Is Not Your Reputation
LinkedIn article commentary
Two physicians can both have a 4.9-star rating and have completely different reputations.
One may be known for catching what others missed and explaining exactly what she sees. Another may be known for kindness, time, and making patients feel heard.
The number is the same. What patients believe they are choosing is not.
I dug into new research on what review language can tell healthcare marketers that the star rating cannot.
Article 02 • Ready / banked
The Technology Was Visible. The Doctor's Expertise Wasn't. Give it distance from the September AI article/post sequence before publishing. Final tightening can wait until it earns a date.
Article 05 • Ready / banked
The Human Relationship Cannot Become Healthcare’s Premium Tier. This is the most ambitious AI/relationship piece and deserves its own runway rather than an already AI-heavy month. Tighten approximately 10–15% when it is scheduled, not while it is banked.
Article 06 • Ready / banked
Subscription Medicine Changes More Than How Patients Pay. It Changes How They Choose. Non-AI long-form piece with a strong acquisition → choice → retention framework. Tighten approximately 10–15% when it is scheduled.
LinkedIn article commentary / wrapper
Jared Dashevsky’s recent piece on the rise of subscription medicine got me thinking about a different part of the model: what changes when patients are not just paying differently, but choosing differently?
When insurance networks and referral structures have less control over the decision, the question shifts from “Who can I see?” to “Who do I want to pay?”
And in a membership model, patients do not answer that question once. They answer it again every month.
Publishing note: tag Jared Dashevsky in the LinkedIn intro post as the inspiration/source conversation. Do not tag him inside the article itself.
Article 07 • Ready / banked
AI Is Adding a Second Algorithm Between Physicians and Patients. Give it at least ~6 weeks of distance from Article 04 because both reference the August 2026 npj Digital Medicine review study. The theses are distinct: Article 04 is about what review language reveals beyond star ratings; Article 07 is about algorithmic mediation of physician reputation.
LinkedIn wrapper
James A. Gardner recently published some fascinating early research on what he calls the “algorithmic editor”: the gap between what patients actually say in reviews and what Google chooses to surface.
I was fortunate to contribute to some of the thinking around trust and anchoring in that work.
But his research left me with another question: What happens when patients stop reading those reviews themselves and ask AI instead?
We may be moving toward a world where two algorithms sit between a physician and a prospective patient.
Publishing note: tag James A. Gardner in the main LinkedIn post, not only in a comment.
Article 08 • Ready / banked
The Best New Feature Might Be One Less Vendor. PMM/GTM piece built from Battle of the Support Stack findings. Do not publish immediately around the Sep 2 Support Stack post or Sep 8 research-feedback-loop post; give the proprietary survey content some distance. This is the strongest current long-form PMM/product-strategy piece.
Next publication priority after Oct 14
-
Article 08 — PMM / product strategy / customer insight
-
Article 06 — subscription model / acquisition + retention / patient choice
-
Article 05 — healthcare systems / AI economics / human relationship
Then return to Articles 02 and 07 after the AI/reputation territory has cooled.
Development guidance
Do not develop another patient-trust/AI long-form article yet. The bank already has enough inventory in that territory.
Prioritize new long-form ideas that demonstrate PMM/GTM leadership: product strategy, buyer/user behavior, adoption, physician workflow, market structure, independent-practice economics, positioning, lifecycle, and growth.
Chief Complaint can become a primary source of future operator/PMM material once the work produces a real decision, tension, customer insight, launch result, conversion result, or tradeoff worth analyzing.
Keep READY separate from NEXT. The bank is inventory; the publishing calendar is a portfolio.
Standalone LinkedIn posts
You Can Trust Your Doctor and Still Not Book the Appointment remains a post, not an article. DPC May Be Outgrowing Its “Cheap Healthcare” Marketing remains in the post bank.
Separate long-form reserve
They Loved the Meet-and-Greet. They Still Did Not Sign Up. is complete in its own file, but the Oct 14 article slot belongs to the broader 4.9-star reputation article.
When a piece rests on my own data, I say so. When it rests on someone else’s, I say whose, and whether they funded it.