How Do I Get My Deli Staff to Sell Catering?

The job this scorecard is hired to do

The complaint sounds like a people problem — "my staff won't sell catering" — but it is almost always a measurement problem wearing a people costume. Walk any deli counter at 12:15 and watch what the operation actually rewards. The line is eight deep, the slicer is running, and the only visible performance signal is how fast the queue clears. Nobody is timing how many customers were asked about their office lunch on Friday. Nobody is counting tray quotes. So the associate optimizes for the thing being watched, which is exactly what a rational person does.

The job this scorecard is hired to do is narrow and specific: make the catering behavior *visible* at the same resolution as the speed behavior. That is it. Not culture change, not a motivational poster, not a manager saying "let's push catering this month" at a pre-shift huddle that half the crew misses. Visibility at equal resolution.

Concretely, that means naming the catering outcomes as discrete, countable KPIs rather than one blurry line item called "catering." A workable starting set for a deli counter:

• Catering asks offered — the number of transactions where the associate raised the catering question at all. This is the top of the funnel and the only KPI fully inside the associate's control.
• Quotes given — an ask that converted into an actual priced proposal, verbal or written.
• Trays and platters closed — the revenue event.
• Deposits secured — for larger orders, the commitment that makes a no-show unlikely.
• Party sub and add-on attach — drinks, desserts, utensil packs, and sides attached to a booked tray.
• Sample offers — a physical taste handed across the counter, which is the single most underused catering tool in a deli.
• Lead-time compliance — orders taken with enough runway that the kitchen isn't buried.
• Speed of service — still on the board, still weighted, just no longer the *only* thing on the board.

Notice the split between controllable and outcome KPIs. "Asks offered" is controllable — a part-timer on a slow Tuesday can run their ask count up regardless of what walks through the door. "Trays closed" is an outcome, dependent on traffic mix, season, and luck. Weight the controllables meaningfully (they're coachable and fair), and weight the outcomes for the revenue reality. A common split is roughly 40 percent on controllable catering behaviors, 30 percent on catering outcomes, and 30 percent on speed and accuracy at the counter — but treat that as a starting posture, not a rule. If your catering revenue is currently near zero, tilt harder toward the behaviors for the first two months, because behaviors are what you can actually coach.

Then score each associate 1-to-5 on every line and roll it into a composite: composite = the sum of (weight × level) across every KPI. A person at level 5 on sandwich speed and level 1 on catering asks lands a mediocre composite, and the gap is legible on the page instead of buried in a manager's head. That legibility is the entire mechanism. Nobody argues with a number they can see and influence.

One caution before you build anything: do not launch the scorecard as a surprise. Publish the matrix — every KPI, every weight, every level definition — before a single score is recorded. Staff who see the rules first read the scorecard as fairness. Staff who get scored first and shown the rules second read it as a trap, and you will spend three weeks recovering trust you did not need to spend.

Why "just tell them to sell" fails at the counter

There are structural reasons a deli associate skips the catering ask, and none of them are laziness. Understanding them changes what you build.

The ask is cognitively expensive during a rush. A sandwich is a closed loop: bread, meat, cheese, wrap, ring, next. A catering conversation opens an unbounded loop — how many people, what date, do we do gluten-free, how much is it, can I get it by nine. Under queue pressure, opening an unbounded loop feels like sabotage to the person holding the line. This is why untrained catering pushes collapse the moment lunch hits.

The associate genuinely does not know the answers. Ask a part-timer what a 24-person sandwich tray costs, what the lead time is, and whether you can do a dairy-free option, and you will often get three shrugs. A person who does not know the answers will not start the conversation, because starting it means being visibly unable to finish it. Embarrassment is a far stronger force than a bonus.

There is no artifact to hand over. Verbal catering pitches evaporate. A customer who hears "we do catering" at 12:20 will not remember it at 4 p.m. when their manager asks them to arrange Friday's lunch. A physical card with three packages, three prices, a phone number, and a QR code converts the ask into something that survives the walk to the car.

The reward loop is invisible. A catering order booked on Tuesday might be fulfilled Friday, paid on the corporate account, and reported in a monthly P&L the associate never sees. From the counter, the catering ask feels like it disappears into a void. Speed gets instant feedback — the line shrinks. Catering gets none unless you build it.

So the fixes map one-to-one against the failures. Give them a scripted two-sentence ask short enough to survive a rush. Give them a cheat sheet with the top three packages, prices, lead times, and the three most common dietary substitutions, laminated and taped where they can see it. Give them a physical card to hand over. And close the feedback loop within 48 hours — post booked orders on the board with the name of whoever planted the seed, even when someone else took the final order.

The two-sentence ask matters more than any other single artifact, so write it down and standardize it. Something in this shape: "Are you ever the one who gets stuck ordering lunch for the office? We do trays — here's the card, and we can usually do next-day with a morning call." It takes six seconds, it's askable while wrapping, it produces a physical handoff, and it deliberately names the customer's *pain* — being the person stuck arranging lunch — rather than your product. That framing consistently outperforms "do you know we cater?" because the second question invites a one-word no.

Train it with roleplay, not a memo. Fifteen minutes at pre-shift, two rounds each, one associate plays a customer who's in a hurry and one plays a customer who bites and starts asking questions. The second scenario is the one that scares people, so rehearse it twice as often. Rehearsal converts an unbounded loop into a bounded one, and a bounded loop is one an associate will actually open during a rush.

How it fits the RevOps stack

A deli counter is a revenue operation with a very short sales cycle, and it responds to the same architecture any RevOps team would recognize: capture the signal, route the lead, quote it, close it, fulfill it, and feed the result back into the scorecard that shapes behavior. The only difference is that the "SDR" is holding a knife and the "CRM" might be a clipboard and a shared inbox.

The critical design point is that the catering ask happens at the counter, but the catering *close* usually doesn't. The customer takes a card, thinks about it, and calls or orders online hours later. If your capture layer can't connect that later order back to the associate who planted it, your scorecard will systematically undercount your best people — and they will notice within about two weeks and stop trying. Solve attribution before you launch, even crudely. Associate-initialed cards, a "who told you about us?" field on the order form, or a per-associate code on the QR link all work. Crude and consistent beats elegant and absent.

Everything upstream of the scorecard is capture. Everything downstream is reinforcement. The loop only spins if both halves exist — capture without reinforcement produces data nobody acts on, and reinforcement without capture produces a bonus that pays the wrong people.

A few practical notes on wiring this in a small operation. Your POS probably already supports a catering department or SKU family; use it, because it's the cheapest attribution surface you own. If you run a separate online ordering platform for catering, put a required "how did you hear about us" dropdown on it, listing your associates by first name. If you're on paper, a duplicate-carbon order pad with an initial box does the job for under twenty dollars. Do not build a custom system for this before you've proven the behavior change — the scorecard is the investment, the tooling is the follow-on.

Upstream, the same logic extends past the deli counter into every adjacent counter-service business with a latent large-order line: bakeries with cake and dessert trays, coffee shops with box coffee and breakfast platters, butcher counters with holiday roasts, pizza shops with team and party packages, and grocery floral with event arrangements. The mechanics port cleanly because the failure mode is identical — a high-frequency small transaction crowding out a low-frequency large one. If you operate multiple departments, build the matrix once and swap the KPI names per counter rather than reinventing the model each time.

What it costs to run this, and the ranges to expect

You can build this for zero dollars, and for a single-location deli you probably should. The honest cost breakdown, from cheapest up:

A spreadsheet — free, plus your time. Columns for KPIs, a row per associate, a weights row, and a SUMPRODUCT for the composite. Genuinely sufficient for a team under about fifteen people. The real costs are the two to four hours to build it well and the ongoing discipline to update it weekly. The failure mode is universal and predictable: the sheet goes stale in week five, the scores stop being posted, and the whole program quietly dies. If you go this route, put the weekly update on someone's actual job description with a named day and time, not on "when I get a chance."

Your existing POS reporting — already paid for. Most modern POS platforms can report by department, by SKU family, and often by employee. That covers your outcome KPIs (trays closed, attach rate, average catering ticket) for free. What it will not cover is your controllable KPIs — asks offered and samples handed out — because those aren't transactions. Those need a manual tally: a clicker, a tick sheet by the register, or an honor-system count at shift end. Manual tallies are imperfect and that's acceptable; you're measuring a trend in behavior, not auditing a bank.

Sales scorecard and gamification platforms — typically low tens of dollars per user per month, some by custom quote. Tools in this category build weighted, multi-metric scorecards, push leaderboards to screens and chat, and tie scores to coaching cadences. They're genuinely good at the visibility layer and they automate what your spreadsheet does by hand. The math rarely works for a single deli — per-seat pricing across a twenty-person hourly crew adds up fast against catering revenue that may still be in the low thousands per month. It starts making sense at multi-unit scale, where you're comparing counters against each other and no single manager can maintain twelve spreadsheets.

Commission and incentive-compensation software — generally custom quote, aimed at enterprise. These platforms model and pay multi-component plans accurately at scale, with audit trails. For a deli, this is dramatic overkill. Mentioned only so you can recognize it and skip it; if a vendor pitches you incentive-comp automation for an hourly counter crew, you're being sold up.

The incentive budget itself — the number that actually matters. This is the line people forget to plan. Whatever you pay out has to come from catering gross margin, not from hope. Work it backward: estimate your catering contribution margin per order, decide what fraction you're willing to convert into incentive, and set the payout so the program self-funds. Anchoring the bonus to the composite score rather than to raw catering dollars is what keeps this honest — it prevents an associate from chasing one enormous low-margin order while ignoring the counter, and it prevents you from paying out on volume that didn't actually clear.

Physical materials — low hundreds, one time. Printed catering cards, a laminated cheat sheet per station, a counter-top menu holder, a whiteboard for the weekly scoreboard, sample cups and picks. This is the highest-ROI spend on the list and the one most often skipped, which is backwards. The card is the artifact that survives the customer's walk to the car.

On payout structure: small, frequent, and visible beats large, delayed, and abstract. A modest weekly amount for the top composite score plus a per-tray spiff moves behavior more reliably than a quarterly bonus, because the feedback arrives while the associate still remembers the conversation that earned it. Add a team component so the person who preps trays at 6 a.m. shares in the win — otherwise your kitchen starts resenting every order the counter books, and a resentful kitchen will slow-walk catering into the ground faster than any incentive can rescue it.

How to evaluate what you've built and shortlist what to fix

Give it a full cycle before you judge anything — six to eight weeks, minimum. Behavior change on a counter is slow, and you'll be tempted to declare failure in week two when the numbers haven't moved. Week two is when the *asks* should move. Revenue lags asks by however long your customers' decision cycle is, which for office lunch is typically one to three weeks.

Read the funnel in order, top to bottom, and fix the first place it breaks:

Asks flat. Nothing else matters yet. The problem is script, confidence, or the associate not believing the score is real. Check whether the scorecard is actually posted and current. Check whether anyone has been recognized for a high ask count. Sit at the counter for one lunch and count asks yourself — the gap between the tally sheet and reality tells you whether you have a behavior problem or a reporting problem, and those need opposite fixes.

Asks up, quotes flat. The ask is landing but not converting to a conversation. Usually the ask is too vague ("we do catering") or there's no artifact to hand over. Tighten the script, get the cards printed, and add a specific price anchor — "trays start around X for a dozen" — because a naked "we do catering" gives the customer nothing to react to.

Quotes up, closes flat. Now you're looking at price, lead time, or menu fit. Pull ten lost quotes and call the customers; five minutes each will tell you more than a month of speculation. Common culprits: lead time too long for how people actually plan office lunch, no vegetarian or gluten-free option, or a price that reads high because it's quoted as a total rather than per person. Quoting per person is a small change that reframes the entire number.

Closes up, margin down. Your incentive is aimed at the wrong target. Someone is discounting to hit tray count. Re-weight toward average order value and deposit compliance, and put a floor under discounting.

Everything up, kitchen in revolt. The most common second-order failure and the one that kills programs. You succeeded at demand generation without expanding capacity. Fix it by enforcing lead-time KPIs, capping trays per day per shift, and cutting the kitchen in on the incentive.

On re-weighting: the ability to change weights overnight is the most underrated property of this model. Game day, graduation, the holiday stretch, a competitor closing down the block — each one deserves a different weight distribution, and you can push a new matrix at Monday's pre-shift and have the team re-aimed by Tuesday. Just don't do it so often that nobody knows what's being measured. Monthly is a good rhythm; weekly is churn.

Finally, audit the scorecard itself for gaming. Every measurement system gets gamed, and that's information, not betrayal. If ask counts spike but quotes don't, someone is muttering "we cater" at retreating backs and clicking the counter. Fix it by weighting quotes above asks, spot-checking with your own observation, and — most effectively — making recognition depend on the composite rather than any single line.

Choosing your path: a decision framework

The right build depends on scale, on whether your catering revenue is currently near zero or merely underperforming, and on how much management attention you can commit weekly. Run this before you buy anything.

Two shortlisting rules worth stating plainly. First, build the matrix before you buy any tool — every platform on the market works better against a matrix you've already defined, and most of them will happily sell you a dashboard for KPIs you haven't chosen. Second, decide where the teeth live before you shop. If the teeth are *visibility*, you want scoreboards and recognition. If the teeth are *pay*, you want your composite feeding a payroll or comp process. Most delis need visibility first and pay second, in that order, and buying the pay layer before the visibility layer is the most common way this program gets expensive without getting results.

Related questions

How long before catering sales actually move?

Asks should move within one to two weeks of a published scorecard. Booked revenue lags by however long your customers' planning cycle runs — typically one to three weeks for office lunch, longer for events. Judge the program at six to eight weeks, not two.

Should I pay a flat spiff per tray instead of building a scorecard?

A flat spiff is faster to launch and does work short-term, but it only rewards the close. Associates with less traffic get nothing, sample offers and asks go unmeasured, and discounting creeps in. Use a spiff as one component inside the composite, not as the whole program.

What if my kitchen can't handle more catering volume?

Then cap it deliberately before you launch. Set a trays-per-day ceiling per shift, weight lead-time compliance heavily so orders arrive with runway, and share the incentive with production staff. Demand generation without capacity planning burns out the kitchen and kills the program.

Does this work for part-time and seasonal staff?

Yes, with a simplified matrix. Cut to three or four KPIs — asks offered, samples given, add-on attach, and speed — using the same weighted 1-to-5 logic. Seasonal hires can be scored from day one because the controllable KPIs don't require tenure.

Can the same model run a bakery or coffee counter?

Directly. Swap trays for cake orders or box-coffee packages and the structure is unchanged. Any counter where a high-frequency small transaction crowds out a low-frequency large one has the same failure mode and responds to the same fix.

FAQ

What if my deli staff simply doesn't want to sell catering?

Reluctance is usually a knowledge and confidence gap, not an attitude problem. An associate who can't answer "what does a 24-person tray cost and when do you need it by?" will avoid starting the conversation to avoid being stuck mid-answer. Give them a laminated cheat sheet, a six-second script, and a card to hand over, then score the ask. Willingness tends to follow capability.

How do I credit an associate when the customer orders three days later?

Build attribution before launch, however crudely. Initialed cards, a per-associate code on the QR link, or a "who told you about us?" field on the order form all work. Uncredited seed-planting is the fastest way to lose your best askers — they notice within about two weeks that their effort disappears, and they stop.

Won't my staff see a scorecard as micromanagement?

That depends almost entirely on sequencing. Publish the full matrix — every KPI, weight, and level definition — before recording a single score, and frame it as the visible path to bonus and recognition rather than a compliance check. Scored-first-explained-second reads as a trap; explained-first reads as fairness.

How much of the bonus should ride on catering versus counter speed?

There's no universal split, and the right one depends on your current gap. If catering revenue is near zero, tilt heavily toward controllable catering behaviors for the first two months because those are what you can coach. As bookings build, shift weight toward outcomes like trays closed and average order value. Re-weight monthly, not weekly.

Do I need software, or is a spreadsheet enough?

For one location and under about fifteen associates, a spreadsheet with a SUMPRODUCT composite is genuinely enough. The binding constraint is discipline, not tooling — name an owner and a weekly update slot. Paid scorecard platforms start earning their per-seat cost at multi-unit scale, when you're comparing counters and no single manager can maintain a dozen sheets.

What's the single highest-return change if I only do one thing?

Print a catering card with three packages, per-person pricing, lead time, and a phone number, and require it be handed to a set number of customers per shift. It's cheap, it converts a forgettable verbal pitch into an artifact that survives the customer's walk to the car, and it gives you a countable behavior to score immediately.

Sources

• https://www.sba.gov/business-guide/manage-your-business/marketing-sales
• https://www.nrn.com/
• https://restaurant.org/research-and-media/research/
• https://www.fda.gov/food/retail-food-protection/retail-food-industry-regulatory-assistance-training
• https://www.dol.gov/agencies/whd/flsa
• https://hbr.org/2015/04/the-problem-with-financial-incentives-and-what-to-do-about-it
• https://www.gallup.com/workplace/236927/employee-engagement-drives-growth.aspx
• https://www.score.org/resource/business-planning-financial-statements-template-gallery
• https://www.nass.usda.gov/

Related on PULSE

• [How Do I Know Where, When, and How Many People to Schedule at Each of My Multi-Unit Retail Locations?](/knowledge/tl0001)
• [How Do I Figure Out How Many People to Schedule Each Day and at What Times for My Single Store?](/knowledge/tl0002)
• [How Do I Know How Many Cooks and Servers to Schedule Each Shift at My Pizza Restaurant?](/knowledge/tl0003)
• [How Many Salespeople Should I Schedule Each Day on My Furniture Store Floor?](/knowledge/tl0004)
• [How Do I Decide How Many Reps to Schedule at Each Store in My Mattress Retail Chain?](/knowledge/tl0005)