He's holding the nine-hundred-pound one
He's been in twice. He's decided on the top-of-the-range machine, he's got his card out, and he has just explained — pleasantly, in passing — that he'll be using it about twice a month to make coffee for himself.
The associate knows the £430 model does that identically. She also knows the difference between those two numbers is a meaningful part of her day's figures, that her area manager reviews average transaction value on Mondays, and that he is not asking for her opinion. He's asking where to tap.
Nothing has gone wrong. There is no complaint, no conflict, no policy. And the next fifteen seconds are one of the hardest things on a shop floor.
The scenario where the skill costs you money today
Every other selling conversation in this series has the associate and the customer pointed roughly the same way. This one doesn't, and the conflict is the reverse of the usual one.
Three features define it.
The customer is satisfied and wrong. He isn't confused or pressured. He has researched it, decided it, and reached a conclusion that doesn't fit what he's just told you he does. Correcting someone who is happy is socially much harder than helping someone who is stuck.
The honest answer has a price, and the associate pays it. Average transaction value, attachment rate, a daily total on a board in the stockroom. Honest product advice in retail is free to talk about and expensive to give, and the person giving it absorbs the cost personally.
And you might be wrong. He may want the expensive one because he wants it. Plenty of purchases aren't optimisations, and an associate who treats every premium choice as an error to be corrected is not being honest, she's being patronising.
That last point is why this is a judgement rather than a rule, and why it needs rehearsing rather than scripting.
What to say
1. Say it early or don't say it. Before the card comes out. Raising a doubt after payment has started reads as an obstacle rather than advice, and it's the version customers find irritating.
2. Ask one more question first. "Can I check what you'll mostly be making?" Sometimes the answer closes it — there's a use case you didn't know about and he's right. Asking costs eight seconds and prevents the patronising version.
3. Say it plainly, once, with the reason. "I'd be doing you out of five hundred quid if I didn't say this — for what you've described, the four-thirty does exactly the same thing. The difference is the milk system, and you'd use it about twice a year." Specific, comparative, no hedging.
4. Give him the version where he's still right. "If you want the bigger one because you want it, that's completely fair and it's a better machine." This is the move that makes the advice land rather than sting. It leaves him a way to buy the expensive one without having been wrong.
5. Then stop. One statement. If he still wants it, sell it to him properly, warmly, with no trace of having been overruled. Repeating the point is no longer advice.
6. Tell him what to do if it's wrong. Return window, what's covered, how to get help setting it up. It's the part that makes the whole exchange read as service rather than a sales tactic.
The half the operator owns
None of this survives contact with the wrong incentive.
If an associate's number drops every time she moves someone from £900 to £430, the behaviour disappears within a month — not because anyone decided it should, but because the scoreboard decided for them. And it's worse than neutral: the store has effectively instructed its staff to let customers buy the wrong thing, without ever saying so.
The fixes are ordinary and mostly absent. Recognise downsell where the customer was better served — a note on the transaction is enough to make it visible. Track return rates and satisfaction alongside transaction value, so the £900 machine that comes back in three weeks shows up against the sale that produced it. And say it out loud, from someone senior, that no one is marked down for talking a customer into the cheaper product.
Until one of those exists, this is a post about a skill nobody is permitted to use.
Four ways it goes wrong
The order-taker, who says nothing. Easiest, most common, and it produces the return, the review, and the customer who doesn't come back.
The late raiser, who mentions it once payment has begun, so honest advice arrives in the shape of an obstruction.
The over-corrector, who assumes every premium choice is a mistake and argues with someone who simply wants the nice one.
The sulker, who gives the advice, loses the sale, and lets it show for the rest of the transaction — undoing the goodwill the honesty just earned.
Why this isn't trained
Sales training is directional. Up, across, attach. Almost no curriculum contains a module on recommending less, so the behaviour has no name, no script and no place on a scorecard.
The incentive points the other way and nobody says so. Everyone on the floor understands the trade-off. It's simply never acknowledged in the training room, which leaves each associate resolving it privately and inconsistently.
Product knowledge is taught as features, not fit. Associates learn what the machine does. Working out which customer needs which model — and therefore who doesn't need the top one — is a different competence and rarely taught.
And peer role play produces a grateful customer. A colleague being talked down from an expensive purchase says "oh, good point, thanks." A real one is mildly embarrassed, or defensive, or has told his partner he's buying the good one — and navigating that is the skill.
What honest-recommendation training can rehearse
A simulation can run the case where the commercially optimal answer and the correct answer differ, and score the correct one — which is exactly the behaviour a live scorecard punishes. Foretell AI supplies the counterparty configuration, transcripts and rubric-based scoring; the product range, margin structure and recognition policy stay with the retailer.
Four to build:
- The over-specified buyer, who has researched his way to the wrong product and is pleased with himself.
- The one who actually wants it, testing whether the associate can accept a premium choice without correcting it.
- The defensive one, who hears the advice as being told he’s got it wrong — the hardest register and the most common.
- The under-buyer, who is about to purchase something that won’t do the job. Same skill, opposite direction, and the one associates find easier because the answer is also the bigger sale.
That last persona matters: without it the module teaches "recommend cheaper," which is just as wrong as the thing it's correcting. The skill is fit, not frugality.
Designing the module
Pass one — the intervention. Score whether a clarifying question was asked, whether the advice came before payment, and whether a reason was given.
Pass two — the exit. Score whether the customer was left a way to buy the expensive one without having been wrong, and whether the point was repeated.
Pass three — the aftermath. Score whether the associate completed the sale warmly either way.
Rubric on observable behavior: Was a use question asked? Did the advice arrive before payment began? Was a specific comparison given? Was the premium choice left legitimate? How many times was the point made? Was the transaction completed without visible disappointment?
Timing is the cleanest measure — advice before the card is service, advice after it is friction, and the difference is entirely in when.
The operator case
Wrong-fit purchases return. The over-specified sale comes back, and it comes back having consumed a delivery, a refund, restocking and often the customer relationship. The transaction value was recognised on the day and the cost lands in a different report.
Advice is the reason to be in the shop. It's the one thing the store has that a browser comparing prices on a phone doesn't, and it only works if it's occasionally against the store's immediate interest. Customers can tell the difference and they calibrate future trust on it.
Transaction-value targets without a quality counterweight buy revenue on credit. Average basket goes up, returns and repeat rate move quietly in the other direction, and the two numbers are usually owned by different people.
And the permission is the whole intervention. Associates already know when someone's buying the wrong thing. What they don't have is a statement from above that saying so is wanted. That statement is free.
For retail and customer experience programmes, this is a useful case in short- versus long-horizon measurement: the honest recommendation is visibly costly today and its return arrives as an absence — the refund that didn't happen — which no standard report can show.
Frequently asked questions
Should a retail associate talk a customer out of an expensive purchase? When what the customer has described doesn't match what the product does, yes — once, early, with a specific comparison, and leaving them free to buy it anyway. Not as a general habit of steering people to cheaper items.
How do you recommend a cheaper product without losing the sale? Ask what they'll actually use it for, give the comparison plainly, and explicitly legitimise the premium choice — "if you want the bigger one because you want it, that's fair." Many customers buy the expensive one anyway and trust the store more for it.
What if the customer buys the expensive one anyway? Sell it warmly and completely, cover the return window and setup, and show no sign of having been overruled. Visible disappointment undoes the goodwill the honesty earned.
How can retailers encourage honest product advice? By removing the penalty. Recognise well-judged downsell, measure returns and repeat purchase alongside transaction value, and state plainly that nobody is marked down for it. Without that, the scorecard overrules the training.
The short version
He's happy, he's decided, and he's about to spend five hundred pounds more than his own description of his life requires. Nothing is wrong, and that's what makes it hard.
Ask what he'll use it for. Say it once, with the comparison and the reason. Leave him a way to buy the expensive one anyway. Then finish the sale properly, whichever way it goes.
And if her Monday number drops every time she does that, she'll stop — which is a decision the store made, not one she did.
Foretell AI lets retailers build conversational simulations — including fit-based recommendation, downsell, and honest-advice conversations like the one above — with configurable counterparties, transcripts, recordings, and rubric-based evaluation. If your average transaction value has no quality counterweight, we're happy to walk through how other operators have structured it.