I got in the car, the driver glanced in the rearview mirror and said: "Hi, I'm Karim. Just confirming you're going to the airport, right? It's about 25 minutes with current traffic. Let me know if you need the AC adjusted."

Four pieces of information. Thirty seconds. Zero friction.

I sat back and thought: this is better onboarding than at least sixty percent of the software products I've used this year.

What Karim Actually Did

Let's break this down properly because there's more happening here than it looks.

He introduced himself. Name, immediately. This is a trust signal. It makes the interaction personal and accountable. Most SaaS products don't do this. You create an account and get dropped into a dashboard with no sense of who is behind the product or why they built it. A name at the start of an experience costs nothing and changes everything.

He confirmed the destination. Before moving, he verified that we both agreed on where we were going. This is the equivalent of a summary screen before a key action. The "are you sure?" that doesn't feel annoying because it's framed as helpful rather than defensive. He wasn't second-guessing me. He was aligning with me.

He set expectations on time. Twenty-five minutes. Not "we'll be there soon." A specific number with a reason attached. Research from Nielsen Norman Group has shown consistently that users tolerate waiting significantly better when they know how long the wait will be. This applies to loading states, onboarding progress, form completion, everything. Karim applied it to a car ride without thinking twice about it.

He gave me control. AC adjustment. A small thing. But offering it signals: you are in charge of your experience here, and adjustments are welcome. How many SaaS products make it genuinely easy to tell them something isn't working for you? How many products even ask?

Four steps. Thirty seconds. The entire experience was set up correctly before we'd gone a block.

Where Most SaaS Onboarding Falls Apart

I'm being a little theatrical about this, but the point is real.

Studies consistently show that the first few minutes of a user's experience with a product determine how they feel about it long after. First impressions in software are sticky. A confusing start creates a skepticism that lingers even after the user figures out how things work.

Most onboarding fails for one of three reasons.

It doesn't confirm the destination. The user signed up for a reason. They have a specific goal in mind. Good onboarding immediately reflects that goal back at them: "You're here to do X. Here's how we'll get you there." Bad onboarding dumps the user into a generic dashboard and says figure it out.

It doesn't set time expectations. How long will setup take? How many steps are there? When will I see my first result? These questions are almost never answered up front. Progress indicators and clear step counts reduce anxiety and improve completion rates. Karim gave me a number immediately. Most onboarding flows give you nothing.

It doesn't hand control to the user. Offering the AC adjustment is the equivalent of asking, "Is there anything about this default experience that doesn't work for you?" at the start. It signals that the experience is configurable and that feedback is welcome. Most onboarding flows are linear and unresponsive. You go through the steps the way the product decided you should, in the order they decided, at the pace they set.

Applying This to Complex SaaS With Multi-Step Setup

The obvious objection at this point is: a car ride is two people and four minutes. My product has fourteen integration steps, three user roles, and an IT approval process. Karim's framework doesn't scale.

It actually does. The principles don't simplify the complexity. They give you a way to carry a user through it without losing them.

The key shift in complex SaaS onboarding is to separate the setup journey from the value moment. Most enterprise products make the mistake of treating them as the same thing. The user has to complete everything before they experience anything useful, which means they're investing significant effort before they have any evidence that the investment is worthwhile.

The Karim model applied to a complex product looks like this. Before the first step, tell the user exactly what the full journey involves. Not a vague "get started" screen. A specific breakdown: "Setup has four stages. This will take about 20 minutes. You'll be able to invite your team after stage two and start using the core feature after stage three." That's the destination confirmation and the time estimate delivered together, before anything else.

Then identify the earliest moment a user can experience genuine value and get them there as fast as possible, even if the full setup isn't complete. Slack lets you send a message before your workspace is fully configured. Notion lets you create a page before you've set up a team structure. The full setup can continue in the background or in subsequent sessions. The user has already felt the product work.

The AC adjustment equivalent for complex SaaS is role-based onboarding. Not one linear flow for everyone. A brief upfront question: "Are you setting this up for yourself or for a team? Are you the technical owner or an end user?" And then a path that's calibrated to what that specific person actually needs. Most enterprise products skip this and send every user through the same flow regardless of their role, which means the admin gets a tutorial about features they'll delegate, and the end user gets a setup screen they don't have permission to touch.

Products Already Getting This Right

The principles aren't new. Some products have been applied well for long enough that the results are visible.

Intercom opens onboarding with a direct question about what the user is trying to accomplish: support, marketing, or product. The entire subsequent experience is tailored to that answer. This is Karim confirming the destination before the car moves. You never feel like you're going through generic steps, because the product has already told you it knows why you're here.

Linear is one of the cleanest examples of human voice in technical onboarding. The setup flow is written in a tone that feels like someone is walking you through it, not like a legal document or a feature checklist. "Let's get your workspace ready" rather than "Complete configuration." The language makes the product feel approachable before the user has done anything. This is the name introduction: making the product feel human before trust has been established.

Figma handles the time expectation brilliantly with its template-first onboarding. Instead of asking new users to build something from scratch, it immediately drops them into a pre-populated workspace. The implicit message is: "You don't need to do anything to see value. It's already here." It removes the blank-canvas anxiety that kills early engagement with creative tools.

Notion gives control from the very first session by letting users choose their starting template based on how they plan to use the product. It's the AC adjustment: before you commit to anything, the product asks which default would work best for you. Users who make an early configuration choice are significantly more likely to continue using a product because they have a sense of ownership over the setup.

The pattern across all of these is the same. They answer the four questions Karim answered, just through product design rather than conversation.

Introducing the Human Element in Enterprise and B2B Products

Enterprise and B2B products have a specific additional challenge: the person doing the setup is often not the person who made the purchase decision, and they may have been handed the product without much context about why the company bought it or what it's supposed to help them accomplish.

This is actually the place where the human element matters most and where it's used least.

A few practical ways to introduce it without a full redesign.

A founder or team note on the first screen. Not marketing copy. A single genuine sentence from the person who built the product, explaining what problem it was built to solve and what the team hopes it does for the user. "We built this because we were tired of X. We hope it means you never have to deal with X again." This sounds small. In an enterprise context where everything feels corporate and impersonal, it lands differently than you'd expect.

Named support contacts during onboarding. Rather than "contact support," show a specific name and photo of the person assigned to onboard new accounts in this category. "Your onboarding contact is Marcus. He'll be checking in at day three." Enterprise users are making significant investments. The sense that a specific person is paying attention significantly reduces the anxiety of a complex setup process.

Progress emails with real context. Most enterprise products send automated setup reminders that read like automated setup reminders. An email that says "You completed stages one and two. Stage three takes about ten minutes and unlocks the reporting dashboard you mentioned in your intake call" does everything Karim did: confirming where the user is, setting expectations, and acknowledging their specific goal.

Explicit permission to go slowly. Enterprise users are often juggling the setup of a new tool alongside their regular work. An onboarding screen that says "You don't have to do this in one session. We'll save your progress and remind you where you left off" removes a significant amount of pressure that most products inadvertently create by making the setup feel like it must be completed immediately.

Measuring Whether Any of This Actually Works

Good intentions in onboarding design are worth nothing if you can't tell whether they're producing better outcomes. The metrics that matter most are not the ones most teams track.

Time to first value is the most important onboarding metric and the most underused. Define what "first value" means for your product, the moment a user has accomplished something real, not just completed a setup step, and measure how long it takes new users to get there. Improvements to your onboarding that reduce this number are improvements that matter. Ones that don't move this number probably don't matter much regardless of how good they feel.

Onboarding completion rate by step shows you exactly where users drop off. Every multi-step onboarding flow has a drop-off cliff somewhere. Most teams know their overall completion rate. Fewer know which specific step causes most of the abandonment, which is the only information that tells you where to actually focus improvement work.

Day-one and day-seven return rates are the clearest signal of whether onboarding set the right expectations. A user who returns on day seven had a first experience that gave them something worth coming back to. A user who doesn't return usually had an onboarding experience that confused them, overwhelmed them, or failed to show them value quickly enough. The Karim framework is specifically designed to improve these numbers by reducing early confusion and creating a sense of momentum before the user has invested much.

Qualitative onboarding feedback at the moment of completion is consistently undervalued. A single open text field at the end of setup asking "Was there anything confusing about getting started?" returns a disproportionately useful signal relative to how cheap it is to add. Users who just completed onboarding know exactly what was hard. Ask them while they still remember.

The combination of time to first value, step-level drop-off data, and day-seven return rate gives you a complete picture of onboarding health that most teams currently get from none of those numbers. Pick one to start tracking if you're not tracking any of them. Time to first value tends to be the highest-leverage place to begin.

When the Product Is Regulated or Security-Sensitive

Healthcare, fintech, legal tech, and government-facing SaaS products face a legitimate constraint that most consumer products don't: parts of the onboarding process are not optional. Identity verification, consent flows, compliance acknowledgements, security setup. These steps exist for real reasons and can't be skipped in the name of a smoother user experience.

The good news is that none of Karim's four principles conflict with regulatory requirements. They just need to be applied around them, not through them.

The most important shift is framing. A compliance step presented as a bureaucratic checkpoint feels like an obstacle. The same step presented as "before we can show you the full platform, we need to verify your identity, this protects your data and takes about two minutes" becomes part of the journey rather than a barrier to it. The step hasn't changed. The user's relationship to it has.

For security-sensitive products, the time expectation principle matters even more than usual. Users in regulated industries are accustomed to verification processes taking unpredictable amounts of time. Being specific rather than vague, "identity verification typically completes within 24 hours, we'll email you the moment it's done," reduces the anxiety that causes support tickets and early churn far more than any visual redesign would.

The destination confirmation also has a specific job in regulated onboarding. The user needs to understand that the friction they're about to experience is temporary and purposeful, and that there's something worth reaching on the other side of it. A fintech product that shows the user exactly what they'll be able to do once verification is complete, with a clear timeline for when they'll get there, transforms a compliance wall into a lobby. The experience hasn't changed. The communication around it has.

Non-Linear and Highly Customizable User Journeys

Not every product has a linear onboarding path and some actively shouldn't. A platform with dozens of features serving five different use cases doesn't have one journey. It has many, and forcing all of them into a single sequential flow is part of what makes so many enterprise products feel disorienting from the start.

The principle that survives non-linearity is the destination confirmation. Even when there's no single right path, there's always a right first question: what are you here to do? The answer to that question should immediately determine which path opens up, which features get highlighted, and which steps can be deferred.

The AC adjustment principle is particularly powerful in non-linear products. Instead of a guided tour that walks every user through every feature in a predetermined order, the most effective approach is to let users tell you what they want to see first. A short upfront question with three or four options, "What do you want to set up today," puts the user in control of the sequence without removing the guidance entirely.

The time expectation needs to be reframed for customizable journeys. Instead of "this will take 20 minutes," the more honest and more useful version is "you can be up and running with your core workflow in about 10 minutes; everything else can be set up when you need it." This gives users permission to make progress without completing everything, which significantly reduces abandonment that occurs when users feel they have to finish a long setup before they can use the product at all.

The key insight is that non-linear products don't need linear onboarding. They need clear entry points, immediate confirmation of the user's current path, and explicit permission to leave and return. Karim's approach on a non-linear journey would be: "You're going downtown, right? We can take the highway or the scenic route, depending on whether you have time. What works for you today?"

Getting Stakeholders to Prioritize Human-Centered Onboarding

This is the conversation most designers dread, and it's worth being honest about why it's hard. Stakeholders who prioritize feature checklists over human-centered onboarding are not being unreasonable. They have evidence that features drive sales, they have a roadmap full of commitments, and they have been told by many well-intentioned designers before you that investing in onboarding experience will pay off, without ever seeing the data to back it up.

The most effective approach is not to argue from principles. It's to show the cost of the current state in numbers they already care about.

Every SaaS product has a trial-to-paid conversion rate. If your onboarding is weak, that number is lower than it should be. A 30-day trial that ends without conversion is a prospect who experienced your onboarding and decided the product wasn't worth paying for. The question you want to put in front of a stakeholder is: what percentage of the people who tried our product and left would have converted if they'd understood its value in the first session? Even a rough estimate of that number, multiplied by the revenue per conversion, tends to make the conversation about onboarding feel less like a nicety and more like a priority.

The second argument is support cost. Poor onboarding generates a predictable category of support tickets: users who can't figure out how to get started, users who set something up incorrectly and need help fixing it, users who completed setup but don't know what to do next. These tickets are expensive to handle and completely preventable. Presenting onboarding improvements as a way to reduce support overhead makes the ROI concrete and connects to a budget line that stakeholders are already motivated to shrink.

The third argument is the competitive one. Have a stakeholder go through the onboarding flow of a direct competitor you know does it well, then go through yours. The experiential gap is often more persuasive than any data point because it's immediate and visceral. People who make product decisions respond to "I felt lost in ours and not in theirs" in a way they don't always respond to percentage points.

Starting From Zero: Practical First Steps for Teams Not Tracking Onboarding

Most teams reading this are not tracking any of the metrics mentioned earlier. That's the real starting point for most product teams, not which metrics to optimize but how to begin measuring at all.

Start with one question and one number.

The question: ask every user who completes onboarding a single open-text question. "Was there anything confusing or frustrating about getting started?" Set it up in whatever survey tool your team already uses. This costs almost nothing to implement and starts generating qualitative signal immediately. You don't need to analyze it systematically at first. Just read the responses for two weeks and notice what keeps coming up.

The number: pick your product's single most important action, the thing that most clearly signals a user has found value, and start tracking how many new users take that action within their first seven days. This is your rough proxy for time to first value before you have the infrastructure to measure it precisely. A week from now you'll have a baseline. A month from now you'll be able to see whether anything you changed moved it.

These two things together, open-text feedback from new users and a weekly active rate for one key action, give you enough signal to make real decisions about where to focus onboarding improvements without requiring a data infrastructure you don't have yet.

The goal is not to build a perfect measurement system before you start improving anything. It's to create a feedback loop that tells you whether the changes you're making are working. Start small, pick one metric, read the feedback, and iterate. The more sophisticated tracking can come once you've established the habit of actually using what you measure.

The Wider Point

Karim didn't go to a UX bootcamp. He didn't read Nielsen Norman Group articles. He just understood, through common sense and probably years of driving, what someone getting into his car needed in the first thirty seconds to feel comfortable and confident in the experience.

That instinct is what good design is trying to replicate at scale. The warmth, the clarity, the brief alignment before the journey starts. None of it requires a complex system. It requires paying attention to what a person actually needs when they first show up somewhere new.

The bar for digital onboarding isn't actually that high. A Tuesday afternoon Uber ride cleared it easily. Most products haven't.

Five stars, Karim. Both for the ride and the lesson.