The First 90 Days: Setting Up a New DSP for Long-Term Success
The decisions made in the first three months determine whether an operation thrives or struggles — and what most new DSP owners get wrong before they realize it.
The routes, the drivers, and the physical setup. Everything built here becomes the baseline the operation runs on for years.
The stand-up, payroll process, communication channel, and first performance data. The operation starts running on systems instead of instinct.
The first real data review, retention check, and the decisions that set the trajectory for month four and beyond. What you do here defines the ceiling.
The first month is spent entirely getting routes running. Nobody builds a stand-up habit. Payroll runs manually and nobody audits it. By day 90, the operation is running — but it's running on guesswork, and the habits that should have been built in week two still don't exist.
The first ninety days of a DSP are unlike any other period in the operation's life. Everything is new — the routes, the drivers, the systems, the relationship with Amazon. Mistakes made in this window cost more than they would later because there's no established foundation to absorb them. A hiring error in month one becomes a retention problem in month three. A route set up incorrectly in week two runs wrong for months.
The decisions that matter most in the first ninety days aren't complicated — they're foundational. The kind of basics that experienced operators do instinctively but new DSPs often skip in the rush to get routes running. Speed matters in the early days, but so does getting the right things in place before the operation has to run at full capacity.
Planning before the routes start — the DSP owners who build lasting operations spend their first thirty days making decisions, not just reacting to them.
Most new DSP owners spend the first ninety days reacting — to driver problems, to route issues, to scorecard flags that could have been prevented. The operations that build lasting success do the opposite. They use the first ninety days to build deliberately: the right hires in the first thirty, the right systems in the second thirty, and the first real review in the third. Each phase builds on the one before it, and skipping any of them quietly creates problems that show up months later when they're expensive to fix.
Here's what each phase looks like when it's done right — and what most new DSPs get wrong in each one.
Hire Right. Get Organized. Build the Physical Foundation.
The first thirty days are the most critical and the most chaotic. Routes are new, drivers are new, everything is being figured out in real time. The temptation is to focus entirely on getting packages delivered and worry about everything else later. That instinct is understandable — but it's exactly what separates the operations that build well from the ones that spend months three through twelve undoing decisions made in month one. The foundation built in these thirty days is the baseline everything else runs on.
The first drivers you hire set the standard — in work ethic, attitude, and what behavior the operation implicitly tolerates. A fast hire who turns out to be unreliable costs more than a careful hire who takes an extra week. Your first cohort of drivers becomes the culture of the operation.
The single most important operational habit in a DSP is the morning stand-up — and the window to build it as a default behavior is day one to day thirty. A DSP that doesn't have a morning stand-up by the end of month one almost never builds one in month two. The habit has to be set while everything else is still being established.
Every driver, every van, every morning — before the first package loads. The pre-trip inspection is the simplest and cheapest accident prevention tool available, and the window to make it non-negotiable is the first week, not the first month. If it starts as optional, it stays optional.
The communication fragmentation problem — messages split across texts, calls, and group chats — starts in week one if you let it. Pick one channel before the first driver starts and make it the only channel for operational communication. It's infinitely easier to establish this as a default than to fix it after six drivers are already using four different methods.
A vacant route feels urgent. But a bad hire in week one becomes a retention problem in week six — and the cost of replacing them far exceeds the cost of covering the route for an extra few days while you find the right person.
The days when the stand-up feels least important are usually the days it matters most. A chaotic morning is exactly when drivers need clear information and a calm, organized start. Skipping it in week two teaches drivers it isn't real.
Manual timekeeping in month one without a review process almost always produces errors. The first payroll cycle sets the driver's expectation of reliability. Getting it wrong on cycle one starts a trust clock that is difficult to reset.
The decisions made in month one — route assignments, pay rates, policies, agreements with drivers — need to be written down. An operation running entirely on memory and verbal agreements in month three is fragile in ways that become very expensive to fix.
Get the first routes running. Run the first stand-up. Establish the communication channel. Do the first pre-trip inspection walk-through with every driver.
Same stand-up format every day. First route calibration based on real timing data. First payroll cycle — audit before it goes out. Fix anything that's running wrong.
One-on-one check-in with every driver. How are they feeling? What's hard? What's unclear? Week three is when drivers decide whether they're staying or looking — don't miss this window.
Review the first month's data. What's working? What needs fixing before month two? Write down the three things that need to change before day 31. Then change them.
The first thirty days are exhausting — there's no way around that. But the operations that come out of month one with a stand-up habit, a communication channel, a payroll process, and a pre-trip inspection routine have built something the chaotic ones haven't: a foundation that scales. Everything added in months two and three goes on top of that foundation. Without it, every new system is just another thing held together by whoever happens to be paying attention that day.
The goal at the end of day thirty isn't perfection — it's intentionality. An operation that made deliberate decisions in month one, documented them, and built the first handful of habits consistently is in a completely different position entering month two than one that just got through it. Both will have survived. Only one of them is actually ready for what comes next.
Before you move into month two, do one thing: write down the three habits you committed to in month one and rate how consistently they ran. Stand-up — did it happen every morning? Pre-trip inspection — is it logged every day? Communication channel — is it actually the only channel? If any of those answers is "mostly" or "sometimes," month two is when you make them absolute. The window to build defaults is closing. Use it.
Build the Habits. Lock In the Processes. Start Tracking.
Month two is where the operation either starts running on systems or keeps running on instinct. The difference isn't visible in the short term — both feel like they're working — but by month six it's unmistakable. An operation built on instinct requires constant attention from whoever knows how things work. One built on systems runs more consistently, catches problems earlier, and doesn't fall apart when a key person is unavailable.
Payroll, rescue, performance tracking, driver check-ins — all four need to be formalized before day 60.
Day 45 is when drivers make their second decision about staying. Miss this check-in and you miss the chance to intervene.
One weekly review session catches patterns before they become expensive — scorecard, telematics, cost per stop.
Thirty minutes to write it down, name an owner, and set a day. That's the difference between a habit and a hope.
By day thirty-one you've run at least one payroll cycle. Now formalize the review — same person, same checklist, same day every week. Every error caught before it goes out is a trust deposit. Every one that gets through is a withdrawal.
By day thirty-one you've had at least one route that needed help. Turn that reactive scramble into a written protocol — who decides, who gets called, how packages move. Write it down, share it, and practice it before you need it under pressure.
Month two is when you start having enough data to see patterns — routes that run long, drivers with emerging telematics flags, stop counts that are miscalibrated. A simple weekly review catches these patterns before they become expensive. You can't fix what you're not measuring.
Month two is when drivers make their second retention decision — around day forty-five. A brief one-on-one catches the ones quietly disengaging before they've decided to leave. Five minutes, direct, genuine — "how's it going, what's hard, what can we fix?"
Payroll runs because the owner remembers. Rescues happen because the manager knows who to call. Performance issues get caught because someone notices — eventually. Everything works, but everything depends on specific people being present. The moment that changes, things fall through.
Payroll runs because the audit checklist runs every Thursday. Rescues happen because the protocol is written and practiced. Performance issues surface in the weekly data review. Driver problems get caught at day forty-five check-ins, not in an exit interview. The operation doesn't depend on any single person's memory.
Same person, same checklist, same day. A fixed weekly commitment before the run processes.
Three questions answered in writing. Everyone in dispatch has seen it and knows it.
Scorecard, telematics, cost per stop — reviewed once a week, every week without exception.
Five minutes, one-on-one, three questions. Every driver before day sixty.
Same time, same format, no exceptions. If it slipped in month two, recommit now.
No fragmentation. If side channels appeared in month two, close them before month three.
The systems built in month two don't feel dramatic when they're working — that's the point. A payroll audit that catches an error before the driver sees it is invisible. A rescue protocol that covers a route in twenty minutes instead of ninety is just a Tuesday. A check-in that catches a driver considering leaving gives you two weeks to act. None of these are heroic moments. They're just the operation working the way it should.
Before day sixty ends, ask yourself honestly which of the four systems is still running on memory rather than process. Pick the weakest one. Spend thirty minutes turning it into a written habit with a named owner and a fixed day. One system formalized before month three starts is worth more than four systems that "mostly happen."
Review What's Working. Fix What Isn't. Plan the Next Quarter.
Month three is the first time you have enough data to make real decisions. Not instinct, not early impressions — actual numbers from sixty-plus days of operation. This is when you find out whether the foundation held, whether the systems are running as intended, and whether the drivers you hired in month one are the ones you want heading into month four. The review isn't a formality — it's the most valuable work you can do before the operation graduates from startup mode into something that scales.
The first time you have enough data to review patterns instead of reacting to individual incidents.
Below 80% means something in months one or two needs fixing before the next hire cohort starts.
Scorecard, retention, route calibration, systems audit — all four before day 90 ends.
Three specific changes before day 91. Not a long list — three actionable decisions based on what the review shows.
Pull the full ninety-day scorecard and look at the trend across all three months — not just the current week. Which metrics improved month over month? Which ones are flat or declining? A single number at day ninety tells you nothing. The direction of travel across three months tells you everything about whether the operation is heading somewhere or just holding on.
DCR and DNR improving month over month. Safety metrics trending down. Week four measurably better than week one across at least three metrics.
Any metric flat or declining in month three after initial improvement. A single metric in the red zone heading into month four.
The trend is the signal. A total number hides whether things are getting better or worse.
Address one thing completely before spreading attention across four.
Transparency about scorecard performance builds trust and shared accountability.
One specific, measurable scorecard improvement to aim for in the next 30 days.
If it's improving — protect whatever is driving that. If it's flat or declining — that's the first conversation of month four.
Count how many of your original hires are still active at day ninety. That number is your most honest operational metric — more revealing than any scorecard figure. If more than twenty percent of your original cohort has left, something in the first sixty days needs to change before you hire the next group. Every replacement hire costs far more than the retention work that would have kept the original driver in the seat.
80% or more of the original cohort still active at day 90. Drivers who stayed can articulate specifically why — not just "it's fine."
Below 80% retention. Multiple drivers who left in the same time window — that pattern points to a systemic issue, not individual circumstances.
Original hires still active ÷ total original hires × 100. That's the number.
Not the official reason — what they told colleagues on the way out. Ask the drivers who stayed.
The answer tells you what to protect and amplify going into month four.
Apply what you learned to the onboarding process before the next hire starts.
Ask them directly: "what almost made you leave, and what made you stay?" That answer is worth more than any exit interview.
Ninety days of operational data tells you which routes are consistently running over time, which stop counts are unrealistic, and which sequences are causing problems that looked like driver issues but are actually routing issues. This is the first real opportunity to recalibrate routes based on evidence. The routes set up in week one were based on estimates. The routes heading into month four should be based on ninety days of real performance data.
Rescue frequency declining in month three. Routes completing within expected windows. Stop counts matching actual delivery pace.
Same routes being rescued repeatedly. Consistent overtime on specific routes. Stop counts set in week one that haven't been adjusted despite evidence.
That list is your recalibration priority. Repeated rescues on the same route are routing problems, not driver problems.
If a route consistently runs 45 minutes over, the stop count needs to come down — not the driver expectation.
Stop sequence problems that showed up in week two are still there in week twelve. Fix them in the review, not on the fly.
Route difficulty matching to driver capability is a month-four decision — make it deliberately, not by default.
Any route rescued three or more times gets a stop count or sequence review before month four starts. That's the threshold.
Go through each system built in months one and two and rate its consistency honestly. The question isn't "do we have this system in place" — it's "did this run every single time without exception." The gap between those two answers is where month four's work lives. A system that runs eighty percent of the time is not a system — it's a good intention that will fail at the worst possible moment.
Every system has a named owner, a fixed day, and a record of running without exception. No system depends on a single person's memory to function.
Any system described as "mostly" running. Any system that skipped even once. Any system where "whoever has time" is the owner.
Same person, same checklist, before every run. Not "mostly" — every single time.
If dispatchers still called around instead of following the protocol, it isn't a system yet.
No exceptions — not even on quiet days or short weeks. Any skip is a signal.
If side channels appeared in month two, close them before month four or fragmentation becomes permanent.
The one rated Inconsistent gets a new named owner and a fixed day before day 91. One system formalized beats four systems that "mostly happen."
Whichever of the four systems ran inconsistently gets a formal owner and a fixed day before day 91. One system, completely formalized.
Every route rescued more than twice in 90 days gets a stop count or sequence review before month four starts. No exceptions.
Anyone who looks disengaged at day 90 gets a direct conversation before month four — not after they hand in notice.
Day ninety isn't a finish line — it's the first checkpoint. The operations that treat it that way come out of month three with a clear-eyed view of where they are, a short list of what needs to change, and the systems already in place to make those changes stick. The ones that just keep running come out with the same problems from month one, slightly more embedded and significantly more expensive to fix.
Before day ninety ends, write down three numbers: your 90-day retention rate, your most-rescued route, and the system that ran least consistently. Those three numbers are the honest summary of your first ninety days — and they're the starting point for everything that comes next.
The routes, the drivers, the stand-up, the communication channel, and the pre-trip inspection — everything built in month one becomes the baseline the operation runs on.
The payroll audit, the rescue protocol, the weekly data review, and the driver check-ins — the operation stops running on instinct and starts running on systems that work without anyone having to remember.
The first real data review, the retention check, the route calibration, and the systems audit — the decisions made here set the trajectory for everything that comes after the first ninety days.
The First 90 Days Build the Operation That Runs the Next 90
The decisions made in the first ninety days don't just affect month one, two, and three — they become the default behaviors of the operation for years. A stand-up habit built in week one is still running in year three. A payroll system put in place in month two is still protecting driver trust in month twenty-four. A route calibrated correctly at day eighty-five is still being run efficiently the following peak season. The investment made in the first ninety days compounds in a way that no amount of reactive fixing in month twelve can replicate.
Most new DSP owners survive the first ninety days. The ones who build lasting operations do more than survive — they use those ninety days to make deliberate decisions, build real habits, and review honestly. The difference between those two approaches shows up slowly at first and then all at once, usually sometime around month six when one operation starts to scale smoothly and the other starts to crack under the same volume.
What did you get right? What would you do differently? Those two answers are the foundation of your month four plan.
If you're below 80%, identify the single change to onboarding that would have made the biggest difference — and make it before the next hire.
Any system rated Inconsistent gets a new named owner and a fixed day before the next quarter starts. One formalized system per quarter builds an operation that doesn't depend on heroics.
The first ninety days are over faster than they feel. What remains is what you built — the habits your drivers experience every morning, the systems that run without being managed, and the data that tells you whether the operation is improving or just surviving. Build those three things in the first ninety days and everything that comes after gets easier. Skip them and everything that comes after costs more than it should.