How Do I Get My Route Drivers to Upsell on Every Stop?

Why the scorecard beats the usual fixes

Most fleets try three things before they try scoring, and all three decay within a quarter.

The pep talk. A supervisor gathers the depot at 5 a.m., says the company needs everyone pushing the new SKU, and sends the trucks out. Compliance is high for about four days. Then a truck breaks down, a customer's dock is blocked, dispatch starts radioing about the on-time number, and selling is the first thing that gets dropped — because it's the only thing on the list nobody is actually counting. Verbal priorities lose to measured priorities every single time. That isn't a motivation problem, it's an arithmetic problem: the driver knows exactly which number gets them called into the office.

The spiff. Twenty-five dollars for every case of the new product placed this month. This works better than the pep talk, and it's genuinely useful for a short promotional push. But a spiff is a bolt-on, not a system. It teaches drivers that selling is an occasional side-quest with a bounty attached rather than part of the job, and when the spiff ends the behavior ends with it. Worse, a narrow spiff distorts the route: drivers chase the one spiffed SKU and quietly stop suggesting the higher-margin items nobody put a bounty on. You get a spike in one line and a dip everywhere else, and the aggregate is often a wash.

The mandate. "Every stop gets an offer, no exceptions." The problem is verification. Nobody rides along on 40 stops a day, so the mandate becomes self-reported, and self-reported offers inflate immediately. A driver checks a box in the handheld, the box means nothing, and now you have a data field that actively lies to you. Mandates without a measurement path are worse than nothing because they create the appearance of a system.

The weighted matrix works differently because it makes the trade-off explicit rather than pretending it doesn't exist. Selling at a stop costs time — usually 30 to 90 seconds. Under a stops-and-on-time regime, that time is pure risk to the driver, so a rational driver skips it. The matrix says: yes, it costs time, and here is exactly how much that time is worth relative to the clock. A driver at level 5 on on-time and level 1 on lines per stop lands a mediocre composite. So does the talkative driver at level 5 on lines per stop and level 2 on stops per route. Balance wins, and balance is legible.

There is one more alternative worth naming honestly: split the roles. Some distributors pull selling off the truck entirely and hand it to a dedicated account rep or an inside-sales desk, leaving drivers to deliver. This is a legitimate model — it's how a lot of larger beverage and food-service operations run — and if your routes are already at capacity with 45-plus stops and tight delivery windows, adding a sales expectation may simply not fit. The matrix approach assumes selling at the door is achievable and valuable. If a time study says it isn't, fix the route density before you fix the scorecard. Choosing the wrong one of these costs you a year.

Choosing the model that fits your routes

The decision isn't "should drivers sell." It's which mechanism carries the weight: visibility, pay, execution tracking, or role separation. Most fleets need two of the four, and the sequence matters more than the tooling.

Start with a time audit before anything else. Pull two weeks of route data and compute average stop duration, windshield time, and slack against the delivery window. If a typical route finishes 45 minutes early, you have room for a 60-second conversation at every stop and then some. If drivers are routinely running 20 minutes late on the last three stops, adding a sales expectation on top of an over-packed route will just make on-time worse and teach everyone the scorecard is disconnected from reality. Reduce stop counts or re-sequence first.

Then decide where the teeth live. Visibility-first — published leaderboards, depot scoreboards, weekly composite rankings — works well on teams under about 15 drivers who see each other every morning and where social proof does real work. Pay-first — commission components on lines per stop, placements, and retention — is the stronger lever when the fleet is large enough that nobody knows everyone's name, or when driver turnover is high enough that peer dynamics never stabilize. Execution-first — handheld prompts, per-stop suggested-order screens, and placement tracking in the field app — is where you go when the constraint is that drivers genuinely don't know what to offer at a given stop.

That last one is underrated. A huge share of "drivers won't sell" turns out to be "drivers don't know what this account already carries, what they dropped last quarter, or what similar accounts nearby are buying." Push a two-line suggestion onto the handheld at each stop — "this account hasn't ordered the 12-oz variety pack in 6 weeks" — and the offer rate moves without any change to comp at all. Data beats exhortation.

A caution on tool selection: buying a gamification or comp platform before the matrix exists is the most common wasted spend in this whole exercise. Every one of those products asks you to define metrics and weights during onboarding. If you haven't already argued that out between operations and sales, you'll end up with defaults that mean nothing, and the platform becomes an expensive scoreboard for numbers nobody agreed on. Build the matrix in a spreadsheet or a free scorecard tool, run it manually for a month, then buy the automation once you know which lines actually move.

What it costs, how long it takes, and what to expect

Direct software cost is the smallest line item. Scorecard and gamification platforms generally land in the low-to-mid tens of dollars per user per month; comp-automation tools are usually custom-quoted and scale with headcount and plan complexity; a spreadsheet is free. For a 30-driver fleet, you're realistically looking at a modest recurring software cost — which is dwarfed by everything else on this list.

The real cost is time and comp redesign. Budget roughly:

• Two to four weeks to define KPIs and weights. This is the hard part, and it's hard because it's a negotiation, not a configuration task. Operations will want on-time and stops weighted heavily. Sales will want lines per stop and placements heavily. Both are right. Expect three or four working sessions before the weights stop moving.
• One to two weeks to validate the data. Every KPI on the matrix needs a source you trust. Offers made is the dangerous one — if it's self-reported with no corroboration, it will inflate. Tie it to something observable: a suggested-order screen the driver has to dismiss, an order line that changed, a callback survey on a sample of stops.
• One month of parallel running before anything is tied to pay. Score everyone, publish the scores, change nothing about compensation. This surfaces broken definitions cheaply and lets drivers see where they stand before the stakes attach.
• One full comp cycle — a quarter in most fleets — before the behavior is settled.

So: roughly 60 to 90 days from first meeting to a matrix that's actually driving pay, if nobody stalls.

On expected impact, be careful about promised numbers. What's defensible is directional. The mechanism that moves is offer rate — the percentage of stops where an add-on was actually suggested — and it moves first, usually within the first two or three weeks of publishing scores, because it's the behavior most under the driver's direct control. Lines per stop follows offer rate with a lag, since not every offer converts. Revenue per route follows lines per stop with a further lag, and it's the noisiest of the three because it's contaminated by seasonality, pricing, and account mix. Measure all three, but judge the program on offer rate early and on revenue per route only after a full quarter.

Watch the counter-metrics just as hard. If on-time percentage drops more than a point or two, or if stops per route falls, the weights are wrong — selling is eating the schedule instead of fitting inside it. If order accuracy slips, drivers are rushing the delivery to buy conversation time, which is a bad trade in every business where a short case triggers a credit and a callback. And watch returns and credits: a driver pushing product onto an account that can't move it generates a return three weeks later, and returns are far more expensive than the incremental case was worth. Cap that by weighting retention and sell-through, not just placement.

One adjacent effect that surprises people: this program usually improves the CRM data more than it improves revenue in the first quarter. Once offers and placements are scored, drivers start recording what happened at the stop with real care, and the account-level picture sharpens. That's a RevOps win in its own right — the sales team suddenly has current visibility into what each account is actually carrying, which changes how the inside desk prioritizes calls. Don't discount it just because it wasn't the headline goal.

Rolling it out and keeping it honest

Implementation fails in predictable places, so run it in a fixed order.

Write level definitions before you write weights. A KPI with a weight but no level definitions is a number nobody trusts. "Level 4 on lines per stop" has to mean something concrete — a specific range, benchmarked against your own fleet's current distribution, not an aspiration. Pull last quarter's actuals, look at the spread, and set level 3 at roughly the median so half the fleet starts near the middle. If level 3 is set at a number only your top driver hits, you've built a scoreboard that says everyone is failing, and drivers disengage from scoreboards that say that.

Publish the whole matrix, weights included. Not a summary, not a rank — the actual sheet. Every driver should be able to compute their own composite and tell you which single line moves it most. A scorecard the driver can't reverse-engineer reads as a black box handed down from the office, and black boxes get gamed rather than met.

Separate the coaching conversation from the pay conversation. The composite drives pay quarterly. The composite drives coaching weekly, and weekly coaching should be about one line, not the whole sheet. "Your offer rate is 22%, fleet median is 41%, let's ride along Thursday and work the first three stops together" is a coaching conversation. "Your composite is 3.1" is not.

Ride-alongs are the honesty check. Once a month, a supervisor rides a full route and independently logs offers made. Compare that against the handheld data for the same day. If they diverge by more than a small margin, your offer-tracking method is broken and needs fixing before the number is worth anything. This is the single control that keeps self-reported metrics from rotting.

Handoffs are where this leaks. Three groups touch the same driver: dispatch owns the route, the supervisor owns coaching, and sales owns the product push. If they grade off different sheets, the driver gets contradictory instructions and defaults to whoever yelled most recently. The fix is procedural — one weekly meeting where all three look at the same composite, and one rule that only the matrix owner changes weights. When a sales manager can informally add a priority outside the matrix, the matrix is dead within a month.

Re-weighting is the feature, not the exception. A promo drops, a new SKU launches, a competitor takes shelf space at a key chain — you shift the weights, republish before load-out, and the fleet re-aims the next morning. But cap the frequency. Weights that change weekly stop functioning as a target and start functioning as noise. Monthly is a reasonable floor for routine adjustment, with the ability to move faster when something genuinely urgent lands.

Applies well beyond DSD. The same structure works on uniform and linen routes (add-on garments, mat placements), pest-control and lawn routes (service upgrades at the door), propane and fuel delivery (tank upsizing, service plans), and parts distribution (adjacent SKU attach). Anywhere a person in a truck has a recurring, trusted relationship with an account, the same two failure modes appear: pure-speed measurement kills the sale, and pure-sales pressure kills the schedule. The weighted composite is the answer in every one of them.

Related questions

Should the upsell target be per stop or per route?

Per route, measured as offer rate across stops. A per-stop mandate is unverifiable and invites false reporting. A route-level percentage lets drivers skip the two accounts where an offer is clearly wrong while still holding a real bar.

What if the account owner is never there?

Score offers to whoever receives the delivery, and add a separate line for accounts flagged as decision-maker-absent. Those get routed to the inside-sales desk instead. Punishing a driver for an absent owner just teaches them the metric is unfair.

Does this apply to union fleets?

The scoring does; the pay linkage may not. Many collective agreements restrict incentive pay structures. Check the contract before designing comp components, and lean on visibility and coaching as the primary teeth if pay changes are constrained.

How do I track offers without adding driver paperwork?

Use the handheld's existing order screen. A suggested-item prompt the driver must accept or decline creates the data as a byproduct of the order they were already entering. Any method requiring a separate log will be skipped.

FAQ

How many KPIs belong on the matrix?

Five to eight. Fewer than five oversimplifies the job and lets drivers optimize one number. More than eight and no driver can hold the sheet in their head, which defeats the purpose. A solid default set: on-time delivery, stops per route, order accuracy, offer rate, lines per stop, new-item placement, and account retention. Each needs a defined 1-to-5 band.

Drivers say selling isn't in their job description. Now what?

Sometimes that's literally true, and you should check. If it is, the conversation is a role-change conversation with a pay change attached, not a scorecard rollout. If it isn't — if the job description already includes account growth — the objection is usually really about fear that sales metrics will penalize them for things outside their control. Publishing the weights and running a month unscored addresses that directly.

What's a realistic offer rate to aim for?

Set the target off your own baseline rather than an outside benchmark, because it varies enormously by route type. Measure the current rate for four weeks first. If the fleet median is 30%, a target of 50% in one quarter is aggressive but reachable. A target of 100% is not credible on any route and signals to drivers that the office doesn't understand the work.

Should the composite affect base pay or just bonus?

Bonus, at least initially. Touching base pay makes the rollout a much heavier lift and raises the stakes on data quality to a level most fleets aren't ready for in year one. A meaningful variable component tied to the composite changes behavior without putting anyone's rent at risk while you're still validating the metrics.

What breaks this program most often?

Weights nobody agreed on. When operations and sales never genuinely negotiated the sheet, each side keeps issuing its own instructions outside the matrix, and drivers learn that the composite doesn't reflect what leadership actually rewards. The second most common killer is an offer-tracking field that inflates and never gets audited.

Does a three-driver operation need this?

The principle yes, the machinery no. A small operation can hold the same weighted view in a weekly huddle and a shared spreadsheet without publishing anything formal. The value is still the same — the owner stops grading only on speed. Automated scoring becomes necessary somewhere north of 10 to 15 drivers, when nobody can hold the whole picture personally.

Sources

• https://www.bls.gov/ooh/transportation-and-material-moving/delivery-truck-drivers-and-driver-sales-workers.htm
• https://hbr.org/2015/04/engaging-your-sales-force-with-the-right-incentives
• https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights
• https://www.salesforce.com/resources/
• https://www.gartner.com/en/sales
• https://www.shrm.org/topics-tools/topics/compensation
• https://www.dol.gov/agencies/whd/fact-sheets
• https://sloanreview.mit.edu/topic/data-analytics-and-ai/

Related on PULSE

• [How Many Sales Reps Do I Need to Hire for My Logistics Company?](/knowledge/tl0058)
• [How Do I Figure Out How Many People to Schedule Each Day and at What Times for My Single Store?](/knowledge/tl0002)
• [How Many Attendants Should I Schedule Each Day at My Car Wash?](/knowledge/tl0067)
• [How Many Producers Do I Need to Hire for My Insurance Agency to Grow My Book?](/knowledge/tl0015)
• [How Many Salespeople Do I Need to Hire for My Car Dealership?](/knowledge/tl0052)