Include eight items in a call evaluation checklist: the start of the conversation, what the buyer needs to achieve, selection requirements, a suitable offer, responses to doubts, accurate promises, the next step and a professional tone. Give each one a rating supported by evidence from the recording; separately flag mistakes that lead the company to promise something it cannot deliver or leave the buyer without the answer they need.
If you and your head of sales keep arguing during reviews about whether salespeople handle conversations well, you need a shared way to assess their work. Otherwise, impressions of someone’s voice will decide whether a criticism stands. A business owner looks for a checklist to get clear grounds for decisions and put the head of sales in charge of listening to calls regularly.
But a long list of requirements can easily start another argument. Points are deducted for a missing question even though the buyer has already given the answer. A salesperson is praised for arranging a call even though they have promised a deadline the company cannot meet. The owner gets a table, but still has to work out what is going on.
First, you need to agree on what action the salesperson was expected to take. And who changes the rule if it makes the employee get in the way of a purchase.
What should this particular call achieve?
Assess the conversation against the task it began with: understand an initial enquiry, answer a question about a proposal or agree a purchase. You cannot expect every contact to cover the whole sale.
Before the table, four fields are enough: call type, what is already known about the buyer, their current question and the expected outcome of the conversation. Fill them in based on the state of the deal before the call. If you define the purpose after listening, it is easy to find a convenient explanation for a successful conversation and a stricter requirement for an unsuccessful one.
For a first incoming enquiry, the outcome is an understanding of what the buyer needs to achieve and a suitable way to continue with the purchase. For an outgoing cold call, it is establishing whether this conversation is needed at all and whether the person is willing to continue. In a follow-up after a proposal, you need an answer to the remaining question and a decision on what to do next. Repeating the entire presentation to a customer you already know just to tick a box is pointless.
A refusal can also be a normal outcome. If the buyer needs a product the company does not sell, the employee should explain this honestly. Not arranging a meeting here is no reason to lower the rating. A different case is a suitable enquiry where the salesperson failed to establish the requirements and decided for themselves that a sale was impossible.
Call evaluation criteria
Call evaluation criteria should describe an action you can hear the salesperson take and how it relates to the buyer’s question. Below is a starting point your sales team can use: it separates a sufficient result from a requirement to say a particular phrase.
| Criterion | What counts as meeting it | When not to lower the rating automatically |
|---|---|---|
| 1. Opening and reason for contact | The person understands who they are speaking to and why; on an outgoing call, the salesperson checked whether it was convenient to continue | A returning customer recognised the salesperson, so the full introduction was not repeated |
| 2. What the buyer needs to achieve | It is clear what the person wants to get, what for and by when; missing details have been clarified | The buyer described their needs in detail, and the salesperson did not ask the same questions again |
| 3. Selection requirements | The constraints needed to choose an offer have been established: volume, compatibility, how the decision is made or another condition relevant to your sale | The budget was not asked about because the price was fixed and the person needed to confirm availability |
| 4. An offer that meets the need | The salesperson linked a specific feature of the product or service to the stated need and explained a significant limitation | They did not list the full range and benefits that were unrelated to the purchase |
| 5. Response to a doubt | They clarified what was unclear and answered the substance of the question; after answering, they checked whether an obstacle remained | They did not use a memorised technique; there were no objections at all |
| 6. Accuracy of terms and promises | The price, what is included, timescales and commitments match the available information and the employee’s authority; unknowns are stated as unknowns | They took time to check instead of making an immediate promise |
| 7. Continuing or ending | For further contact, both sides know who will do what, when and how they will get in touch; if contact ends, the reason is clear | They did not arrange another call after a final refusal or a question that had been fully resolved |
| 8. Professional tone and attention | The salesperson let the person finish an important point, answered the question, did not apply pressure and was not rude | They spoke without exaggerated enthusiasm, paused and briefly clarified something unclear |
On the form you use, each item needs a rating, a timestamp in the recording and a brief reason. For the item on accurate promises, add a link to the company’s current terms: the conversation alone cannot tell you whether the employee has the authority to give that deadline. Assess the content of the conversation from the recording, and also use the document in force at the time to check whether the promise was accurate.
Do not turn the row on selection requirements into a compulsory interrogation. Write down the information without which your team risks making an unsuitable offer: for equipment, this is compatibility; for a service, what the work includes; for a delivery, the date and volume. Other questions need a separate explanation of why the buyer and salesperson need them.
Use the same logic for comments about price: the feedback must show which question the salesperson failed to answer. You cannot lower the rating just because no discount was offered. What to examine before offering one is covered in the article “The customer said ‘It’s expensive’. And you are already planning to lower the price”.
How do you assign ratings and review disputed items?
Four ratings are enough to start: met, error, not applicable and insufficient data. They preserve the meaning of the feedback, which is lost in a single overall score.
| Rating | Basis | What happens next |
|---|---|---|
| Met | The required action is audible and addresses the task of this conversation | An excerpt of a successful response is saved so it can be reviewed with the team |
| Error | The action was required and the employee had an opportunity to take it, but missed it or did it incorrectly | The excerpt, the consequence for the purchase and the correction needed are recorded |
| Not applicable | The situation that would require the action did not arise in this conversation: for example, the buyer expressed no doubts | The item is excluded from the requirements for this call, with a short explanation |
| Insufficient data | There is not enough audible material or confirmation of the terms to reach a conclusion | The rating for this item is postponed until supporting evidence is available |
A “not applicable” rating does not excuse something the salesperson themselves failed to do. If they began the presentation without understanding the enquiry, they should have found out what the buyer needed. If the customer ended the conversation immediately, there is no presentation to assess. Record the difference in the rule so that each listener does not have to establish it again.
Feedback also follows a clear format: the requirement, an exact quote or the absence of a required action, the timestamp of the excerpt, the consequence and an acceptable response. You cannot invent a quote for a missing question: identify the part where the question was needed. Offer wording for the response as a recommendation, without presenting it as something said in the recording.
For example, when the delivery date is unknown, an acceptable response has three parts: what has already been confirmed, what is still being checked and when the buyer will get an answer. Requiring a salesperson to confidently give a date the company does not have means teaching them to make promises at someone else’s expense.
The head of sales reviews an employee’s disagreement with a rating using the same excerpt and the rule in force at the time. If the reason for the criticism no longer holds, the rating is corrected. The owner should be wary when only the salesperson is allowed to be wrong in the team, while the reviewer is always right.
Which mistakes must not be hidden in an average score?
Flag separately any actions that require intervention in the deal: an unconfirmed promise, misstated terms, rudeness or an ignored question that the buyer needs answered before they can continue. A good greeting does not lessen the consequences of these mistakes.
If your version has ten applicable requirements and nine have been met, the calculation gives 9 ÷ 10 × 100 = 90%. But if the tenth requirement was not met because a promise to dispatch the goods was made before an available date was agreed, that 90% says nothing about whether the promise can be kept. The head of sales needs to find out the actual date and arrange to clarify it with the buyer before they start relying on the promise.
In a manual checklist, this percentage is acceptable as a count of the items met. Do not call it the probability of a sale or an overall assessment of professionalism. Reducing the number of applicable requirements changes the percentage without improving the conversation, so always show the number of requirements and a separate list of critical errors alongside it.
Approve the list of critical errors before assessment begins. For each one, record who takes over the deal and before which event they must intervene: payment, dispatch or the promised answer. This makes the rating a signal to act. Red in a report is not enough on its own.
How do you introduce the checklist in your team?
Start with one type of call and a short trial of the rules. The head of sales collects recordings and leads reviews; the owner approves the key requirements and checks whether they lead to clear decisions.
First, find out whether people understand the criteria in the same way
For the first trial, take five conversations of the same type from different employees: some with a next step, some with a refusal and some with no clear outcome. This is a small set for adjusting the form, not a measure of the quality of the whole team. Two people assess them independently, without seeing each other’s ratings: for example, the head of sales and the employee assigned to check calls.
Compare the disagreements for each item. One person heard a response to a doubt, the other heard the question being avoided: you need a specific excerpt and a clearer definition of what counts as a sufficient answer. After revising the rules, review three new recordings. End the trial when the conclusions match on all critical requirements and the other disagreements can be explained by the recording and the rule. There is no reason to average two team leaders’ opinions here.
Then set the number of calls the team can review
For the first working week, you can take three conversations per salesperson, selected at random within the chosen type. Add enquiries involving complaints or promises that pose a risk separately. You must not mix them into the regular sample when calculating the proportion of errors: you selected problematic conversations in advance.
Work out the time using your own recordings: multiply the number of salespeople by the number of calls and by the actual time spent listening and filling in the form. For example, five salespeople, three calls, ten minutes per review: 5 × 3 × 10 = 150 minutes. Another ten minutes for a conversation with each employee makes 200 minutes in total, or 3 hours 20 minutes a week. If the head of sales does not have that time, the resources for monitoring must be in place before you promise to cover the whole team.
Make sure feedback leads to a change in the conversation
After the review, choose one recurring action to correct. The salesperson receives the excerpt, an explanation of the mistake and a suitable way to respond. In the next five calls where the same task arises, the head of sales records whether the employee took the action on their own. If the relevant situation did not arise, there is not yet enough material for a conclusion.
The owner only needs to see which mistake is being corrected, how many relevant conversations have already been reviewed and in how many it occurred again. For example, it happened again in two out of five: 2 ÷ 5 × 100 = 40% in this small sample. This is a reason to keep working on the specific action, without drawing a conclusion about the proportion of revenue lost.
When several salespeople make the same mistake, examine the requirement itself and the information available to them. If everyone is promising a date based on an outdated table, each person is missing the same document. Training them to speak confidently will not remove that cause.
When is manual call review no longer enough?
Manual reviews no longer fit into the workload when the agreed sample is skipped or feedback arrives after the point when the buyer can still be helped. Then there is a task for automation: find breaches of known rules in the available conversations and give the head of sales supporting evidence for review.
At RENTROP, AI ROP handles this task: sales quality control shows a salesperson’s mistakes with a quote and how they should have responded, taking the company’s rules into account. The service only reads data; putting things right with the buyer and speaking to the employee remain the head of sales’ responsibility.
It makes sense to assess automation against the same set of recordings the team used to agree the criteria: whether it spots the relevant mistake, correctly attributes a remark to the salesperson and avoids requiring an unnecessary action. Otherwise, an argument over the listener’s personal preferences will simply become an argument with the software.
The checklist must stand up to a salesperson’s objections. If the employee skipped an unnecessary question and met the buyer’s need more quickly, the criticism must be withdrawn and the rule clarified. It is at precisely this point that the owner sees what their team is really learning: to help the buyer make a purchase or to guess the reviewer’s requirements.