Dragon-Led Growth: DLG for Modern SaaS
I want to introduce a term: Dragon-Led Growth, or DLG.
This is not supposed to replace PLG, SLG, enterprise sales, consulting, or any other acronym people love to throw around. It is a way to describe what I think modern SaaS is increasingly becoming, especially in the AI era.
The short version:
Dragon-Led Growth = vision-led company + sales-led discovery + service-led deployment + productization loop.
That is the whole concept.
The company needs a strong vision at the center. Not just “we sell workflow automation” or “we make teams more productive”. I mean an actual directional belief about how the world should work. The kind of belief that makes product decisions easier, hiring easier, sales easier, and customer conversations sharper.
Then you go close to the customer. You do not wait for the perfect self-serve funnel to magically explain the market to you. You use sales-led discovery to understand the real pain. You use service-led deployment to make the product work in the customer’s actual environment. And then you productize what repeats.
That is DLG.
Why This Is Modern SaaS
Classic SaaS was built around low marginal cost. Marginal cost simply means the extra cost of serving one more customer. The dream was: build the product once, sell it many times, keep onboarding light, keep services low, and preserve high gross margin.
That model is still great when it fits.
But AI is pushing SaaS into messier workflows. Enterprise AI is not just a login screen and a few settings. It needs customer data, permissions, integrations, trust, evals, change management, security reviews, and workflow redesign. The real value is often behind the simple demo.
So the modern SaaS motion becomes more active:
Vision
-> Customer discovery
-> Hands-on deployment
-> Reusable product learning
-> Stronger platform
-> Bigger vision
This is why Palantir is such an important North Star. Not because every company should copy Palantir exactly, but because Palantir understood something early: if you want to solve truly important enterprise problems, you often need to send technical people into the field. The product does not become weaker because of that if the field learning returns to the platform.
Databricks is another useful North Star. It is a platform company in a very complex domain: data, AI, governance, infrastructure, enterprise adoption. That kind of company does not win only by having a nice UI. It wins by becoming part of the customer’s operating model.
And that is the point of DLG. The goal is not to make a tool. The goal is to build an operating layer.
The Vision-Led Part
The most important part of Dragon-Led Growth is the vision.
Without vision, sales-led discovery becomes “ask customers what they want”. That sounds customer-centric, but it can be dangerous. Customers are great at explaining pain. They are not always great at designing the product category you should build.
So the company has to be stubborn on direction and flexible on details.
That sounds simple, but it is hard. It means you listen deeply to customers, but you do not let every customer become the roadmap. You understand their workflow, but you keep asking: does this fit the mission? Does this make the platform stronger? Does this teach us something reusable?
The best companies have this tension. They are extremely customer-obsessed, but not customer-controlled. They work backwards from real pain, but they still build toward a clear future.
That is the “dragon” part for me. A dragon has direction. It is not a collection of random customer requests with a pricing page attached.
Sales-Led Discovery
Sales-led discovery does not mean “sales makes promises and engineering suffers later”.
That is the bad version.
The good version means sales becomes a discovery engine. Sales conversations are not just qualification calls. They are research. They reveal where the budget is, where the workflow hurts, where the current system fails, who owns the problem, and how urgent the pain is.
In DLG, sales should find the pain that is both:
- important enough that the customer will pay
- repeatable enough that the product can grow from it
That second point is crucial. A huge one-off problem can be attractive, but it can also pull you away from the company you are supposed to build.
Service-Led Deployment
Service-led deployment means you are willing to help the customer get to value.
This can involve forward-deployed engineers, solution architects, implementation engineers, customer engineers, AI deployment teams, or partners. The title does not matter. The function matters.
The function is to cross the gap between “the product works in a demo” and “the product works in the customer’s real workflow”.
This is where AI changes the game. Custom demos are faster. Integration work is faster. Data mapping is faster. Documentation is faster. Testing customer-specific workflows is faster. Productizing repeated patterns is faster.
So yes, the company might accept more implementation work than classical SaaS would like. But the bet is that AI and productization keep this from becoming consulting.
The Productization Loop
This is where DLG either works or fails.
Every customer deployment should leave something behind:
- a reusable connector
- a workflow template
- an eval set
- a prompt pattern
- a migration script
- a permission model
- a demo environment
- a product feature
- a repeatable implementation playbook
If nothing reusable remains, you did a project. Maybe it was profitable, maybe it helped the customer, but it did not compound.
DLG is about compounding.
The simple test:
Did this customer make the next similar customer easier to win and deploy?
If yes, the work is an asset.
If no, it is a tax.
Why It Works
DLG can create a very strong company because it combines several powerful things.
You get the clarity of a vision-led company. Everyone knows what the company is trying to become.
You get the market signal of sales-led discovery. You are close to real budgets and real pain, not just abstract product ideas.
You get the adoption power of service-led deployment. The product actually lands in the customer’s workflow.
You get the economics of SaaS if you productize correctly. The first deployment may be heavy, but the tenth should be lighter, and the hundredth should feel like a platform.
You also get moat. The customer is not just using a tool. The customer is changing workflows around your system. That creates switching cost, expansion surface, cross-sell, upsell, and much better customer understanding.
What To Be Careful About
DLG is powerful, but it is dangerous.
The obvious danger is becoming consulting. If every customer gets a unique solution and nothing feeds back into the platform, you do not have DLG. You have services.
The second danger is weak vision. If the company does not know where it is going, every large customer can pull it somewhere else. Revenue starts to look good, but the product gets worse.
The third danger is margin denial. Yes, it can be smart to trade some gross margin for moat early on. But eventually revenue must scale faster than delivery headcount. If every deal needs the same amount of manual work forever, the model is broken.
The fourth danger is fake customer obsession. Customer obsession does not mean blindly doing what customers ask. It means understanding the underlying pain so well that you can build the better answer.
The DLG Formula
If I had to compress Dragon-Led Growth into one formula, it would be this:
Strong vision
+ close customer discovery
+ hands-on deployment
+ ruthless productization
= modern SaaS compounding
That is the way I would build many AI-native SaaS companies today.
Not every company needs this. Some products should stay pure PLG. Some products should be narrow, self-serve, and simple. That is fine.
But if you are building in a complex enterprise domain, especially with AI, DLG is the model I would take very seriously.
Be vision-led. Get close to the customer. Deploy the product into reality. Productize everything that repeats.
That is Dragon-Led Growth.