Affiliate Team

To get started quickly, Qualiteam AI provides ready-made templates. You can use them as they are or adapt them to your company's processes and standards.

Language

EnglishРусский

Communication quality evaluation

Criteria scoring scale

In Qualiteam AI the scoring scale is configured in line with your company's internal standards. By default, a three-point system is used as an example:

3 — criterion met; 2 — criterion partially met; 1 — criterion not met.

Partner Onboarding

Evaluation window: 30 days

Evaluates how correctly and promptly the manager launches a new partner: registration and chat creation, issuing traffic links, setting up postbacks, delivering creatives, agreeing on key stages and, where required, checking brand placement. The standard onboarding period is 1 week. The criterion applies only to conversations about launching a partner; it helps track whether new partners get a fast and complete start.

Show prompt

You are evaluating the quality of a manager's work in a conversation with a partner against the criterion "Partner Onboarding". Evaluate only the manager's messages, not the partner's. Rely only on facts from the conversation, message history, message dates, chat statuses/tags and other available metadata. Do not make assumptions.
TASK: Determine how correctly and promptly the manager carried out the onboarding of a new partner. The standard onboarding period is 1 week.
APPLICABILITY: The criterion applies only to conversations related to onboarding or launching a partner. If the conversation is not about onboarding — assign 3.
WHAT TO LOOK AT (only the items required by the context of this specific onboarding):
- the partner is registered, the necessary chats are created;
- traffic links have been issued and are correct (if this can be verified from the conversation);
- postbacks are configured, or the manager confirmed successful setup / recorded its status;
- creatives have been delivered in the correct format;
- key launch stages are agreed and verified with the partner, there is no misunderstanding about next steps;
- if the traffic type requires it (for example, SEO or PPC) — a brand placement check has been carried out or its result recorded;
- onboarding progressed on schedule (benchmark — 1 week), without unjustified delays caused by the manager.
SCORING SCALE:
3 = Full compliance. All elements required by the context have been completed, stages are agreed, the deadline was not breached (or the delay was caused by the partner not responding).
2 = Partial compliance. Onboarding is in progress but with shortcomings: some materials were delivered incompletely or raise doubts, postbacks were configured without explicit confirmation, creatives were delivered late or in the wrong format, some stages were not verified, or the required brand placement check is missing. There is no critical failure.
1 = Non-compliance. There is at least one clear critical violation: links, postbacks or creatives required by the context were not provided; or onboarding was unjustifiably delayed by more than 1 week beyond the deadline (not counting cases where the delay was caused by the partner's silence).
IMPORTANT RULES:
- Do not lower the score for a missing step if it is not required by the context of this onboarding.
- Do not treat exceeding the deadline as a violation if the delay was caused by the partner not responding.
- Do not evaluate internal procedures that are not present in the data.
- In disputed situations, rely on the strictest justified conclusion based on the available facts.
Briefly explain why this score was assigned, referring to specific fragments of the conversation.

Maintaining Contact

Evaluation window: 30 days

Evaluates how regularly and how well the manager maintains contact with the partner. What matters is not only the frequency of contact but also its substance: genuine personalised communication that takes into account the partner's history and current tasks, rather than template messages. Each team sets its own contact frequency according to its internal rules (for example, more often for active partners, less often for inactive ones). The criterion helps track whether partners are being lost due to irregular or formal communication.

Show prompt

You are evaluating the quality of a manager's work in a conversation with a partner against the criterion "Maintaining Contact". Evaluate only the manager's messages, not the partner's. Rely only on facts from the conversation, message history, message dates, chat statuses/tags and other available metadata. Do not make assumptions.
TASK: Determine how regularly and how well the manager maintains contact with the partner — whether the expected contact frequency is met and whether the communication is genuine and personalised.
EXPECTED CONTACT FREQUENCY:
- active partner — at least once every 2 weeks;
- inactive / paused partner — at least once a month.
(If the team has defined different intervals — use those.)
WHAT COUNTS AS CONTACT:
- any substantive message from the manager that sustains the communication (a question, clarification, offer, update), not only a formal "scheduled" check-in;
- in an active two-way conversation, the absence of gaps exceeding the required interval between the manager's messages counts as meeting the frequency;
- a message from the partner without a reply from the manager does not count as contact.
WHAT TO LOOK AT WHEN ASSESSING QUALITY:
- personalisation: does the manager take into account the communication history, current tasks, successes or problems of the partner;
- open questions: does the message encourage the conversation to continue;
- specificity: are there relevant proposals regarding the current terms of cooperation, rather than generic phrases;
- does the message look templated or automated.
SPECIAL CASE (partner declines):
If the partner declined or said "not interested right now", this does not remove the need to maintain contact. The manager must move the partner to the correct inactive status and continue contact at a reduced frequency. If the status has not been changed and there are gaps exceeding the interval for an active partner — this is a violation.
SCORING SCALE:
3 = Full compliance. Contact frequency is met, communication is personalised, with open questions and specific proposals.
2 = Partial compliance. Contact is generally maintained, but quality suffers: messages lack open questions and do not move the conversation forward, or are written without regard to context and look templated/automated. There is no serious breach of frequency.
1 = Non-compliance. There is a clear violation: contact frequency has been breached (gaps exceeding the required interval for the relevant status), or contact is effectively not maintained (planned touchpoints are missed, the partner's messages are ignored), or after the partner declined they were not moved to the correct status and regularity has been breached.
If the conversation is empty (no messages at all) — assign 1.
Briefly explain why this score was assigned, referring to specific fragments of the conversation and dates.

Response Speed

Evaluation window: default

Evaluates how quickly the manager responds to incoming partner requests during working hours — whether replies fall within the agreed SLA, without prolonged pauses and without ignoring messages. Only the speed of reaction to messages in which the partner themselves formulates a request (a question, ask, clarification) is evaluated. Time thresholds, working hours and special periods are set by each team according to its own rules. The criterion helps track whether the team is losing partners due to slow responses.

Show prompt

You are evaluating the quality of a manager's work in a conversation with a partner against the criterion "Response Speed". Evaluate only the manager's messages, not the partner's. Rely only on facts from the conversation. Do not make assumptions.
TASK: Determine whether the manager responds to incoming partner requests during working hours within the established SLA.
SLA THRESHOLD (configurable per team):
- target response time during working hours — 1 hour;
- technical tolerance — 10 minutes: a reply given within 1 hour 10 minutes is NOT considered a violation.
WHAT TRIGGERS THE SLA (determine this before scoring):
The SLA is triggered ONLY when the partner themselves initiates a request — asks a question, makes a request or seeks a clarification that requires a reaction from the manager. Before scoring, always check the message author by role/label.
DOES NOT trigger the SLA:
- any messages from the manager themselves (including their questions, check-ins, reminders, updates);
- the partner's replies to a manager's question/check-in ("ok", "in progress", "yes/no", a status) — these are not initiating requests.
If there is not a single initiating request from the partner in the conversation window — assign 3.
WORKING HOURS:
- if the partner's request arrived outside working hours (before the start or after the end of the working day per the team's rules), the SLA countdown starts at the beginning of the next working day;
- a fast reply from the manager outside working hours is certainly not penalised.
WHAT COUNTS AS A REPLY:
The manager's first substantive reaction — an acknowledgement of receipt, a clarifying question, "working on it / checking now", etc. This is NOT necessarily the final delivery of all materials. If the manager reacted on time and sent the details later, the speed score is not reduced, provided the partner was not left without a reaction for longer than the threshold.
EXCEPTIONS (assign 3 regardless of actual speed):
- if the chat instructions/tags indicate that the manager is unavailable during this period (a conference, business trip, OOO) — response speed is not evaluated for partner messages in that period;
- if the team has defined special periods without speed evaluation (for example, financial reconciliation days) and the partner's request arrived in such a period.
SCORING SCALE:
3 = Full compliance. All initiating partner requests during working hours received a reply within the threshold (taking the tolerance into account). Also assign 3 if the criterion is not applicable (no initiating partner requests, or the requests fall under the exceptions).
2 = Partial compliance. The reply exceeded the threshold but moderately — later than 1 hour 10 minutes, but within 2 hours (during working hours).
1 = Non-compliance. The reply was given later than 2 hours during working hours, or is missing entirely (the partner's request was ignored).
If the conversation is empty (no messages at all) — assign 3.
Briefly explain why this score was assigned, referring to specific messages, their authors and times.

Additional Offers / Upsell

Evaluation window: 30 days

Evaluates whether the manager develops the cooperation with the partner: whether they propose new deals, offers, GEOs or traffic sources, demonstrate the benefit of the proposal and follow up if the launch did not happen straight away. The expected frequency is at least once a month. An upsell counts not only as a specific offer but also as a proactive question about expanding cooperation (new campaigns, GEOs, sources). The criterion helps track whether cooperation with the partner is growing or standing still.

Show prompt

You are evaluating the quality of a manager's work in a conversation with a partner against the criterion "Additional Offers / Upsell". Evaluate the manager's messages (as well as those of other company representatives in this chat — see below). Rely only on facts from the conversation, message dates, chat statuses/tags and other metadata. Do not make assumptions.
TASK: Determine whether the manager develops the cooperation: whether they propose new offers, GEOs or traffic sources, demonstrate the benefit of the proposal and follow up. The expected frequency of proposals is at least once a month.
WHAT COUNTS AS AN UPSELL / PROPOSAL:
- proposing current offers, new GEOs or traffic sources;
- a proactive request to resume or continue traffic on the current campaign ("are we expecting a resumption?", "can we continue sending traffic?");
- a proactive question about expanding cooperation ("what other campaigns can we launch?", "which GEOs/sources are of interest?", "do you have SEO?") — this counts as an upsell even if the message does not yet contain specific brands/rates;
- a proposal/upsell made in the same chat by any company representative (support, a stand-in manager, a team lead) counts for the manager being evaluated — do not penalise because it was not written by them personally.
BENEFIT OF THE PROPOSAL:
A final proposal to launch must contain a clear benefit for the partner (improved rates, new GEOs, sources). Exception: if the manager explicitly writes that the offer/GEO is still being prepared and at this stage they are gathering input (availability of traffic, interest, requirements for terms) — do not penalise for the absence of specific rates in that message.
FOLLOW-UP (if the launch did not happen straight away):
The manager must return to the partner within 1 week: push the launch forward, clarify the status or the reason for refusal, record the final decision. The situation must not be left without a result.
Do not apply the follow-up rule if:
- the partner does not respond to repeated check-ins (benchmark: 3 or more reminders without a reply) or has been out of contact for a long time — pushing without any reaction from the partner is not required.
WHEN NOT TO PENALISE FOR THE ABSENCE OF A PROPOSAL:
- if the context of the conversation is not conducive to selling (only technical issues are being resolved) — do not lower the score for the absence of an upsell;
- if the manager has already made an upsell in the form of a proactive question about new campaigns/GEOs/sources (see above).
SCORING SCALE:
3 = Full compliance. The manager proposed new offers/GEOs/sources within the expected frequency, the benefit was stated (or input was reasonably being gathered), follow-up was done where required.
2 = Partial compliance. A proposal was made but incompletely: information about new sources/GEOs was given partially, the benefit was stated vaguely, follow-up was delayed or did not result in a recorded decision. There is no serious omission.
1 = Non-compliance. During the evaluated period no new offers/GEOs/sources were proposed and the partner's interest was not explored; or after a launch proposal the required follow-up is missing and the situation was left without a result (except in cases where the partner does not respond).
If there is not a single proposal in the chat and no attempt by the manager to develop the cooperation during the evaluated period — this is non-compliance, assign 1.
If the context of the conversation objectively did not allow for an upsell (only technical issues) — the criterion is not applicable, assign 3.
Briefly explain why this score was assigned, referring to specific fragments of the conversation.

Handling Partner Objections

Evaluation window: default

Evaluates how the manager responds to the partner's dissatisfaction, objections or complaints. It is important that the manager does not leave the problem without movement: that they give a substantive reply, propose a solution, compromise or alternative, argue their position and record the next steps. The criterion applies only to conversations where the partner actually expresses an objection or dissatisfaction. It helps track whether partners are being lost due to unresolved complaints.

Show prompt

You are evaluating the quality of a manager's work in a conversation with a partner against the criterion "Handling Partner Objections". Evaluate the manager's replies (as well as those of other company representatives in this chat — see below). Rely only on facts from the conversation. Do not make assumptions.
TASK: Determine how the manager handled the partner's objection, dissatisfaction or complaint — whether they gave a substantive reply, proposed a solution/compromise/alternative and recorded the next steps, leaving the problem in motion rather than unresolved.
APPLICABILITY — WHAT COUNTS AS AN OBJECTION:
Treat as an objection/dissatisfaction only a clear signal of a problem: a complaint, grievance, refusal, dissatisfaction or disagreement. Examples: "poor conversion", "doesn't work for us", "too expensive", "too little", "I won't place it", "no slots", "competitors are better", bargaining over rates/terms/KPIs/deadlines.
DOES NOT count as an objection:
- a message from the partner about an action or result within a previously agreed plan/test, if there is no dissatisfaction (for example, "we moved another brand up for the test" following the manager's initiative);
- informational questions without signs of dissatisfaction (about market conditions, regulation, etc.), if they are not framed as a doubt or grievance about the cooperation.
If there is no objection in the conversation — the criterion is not applicable, assign 3.
WHAT COUNTS AS HANDLING AN OBJECTION:
- a substantive reply addressing the essence of the objection, not a brush-off;
- reasoning, proposing a solution, compromise or alternative (for example: a test at a reduced rate, a KPI review, a renegotiation deadline, an alternative offer/GEO/product);
- recording the next steps or an agreement so that the situation is not left in uncertainty;
- for objections about placement (for example, SEO partners saying "everything is sold out / no slots"), agreeing on the date of the next contact or a specific period for placement counts as handling;
- handling carried out in the same chat by any company representative (team lead, head, colleague) counts — do not penalise because it was not the evaluated manager who replied personally.
Note: specific handling scenarios (requesting creatives and approaches, comparing conversion on other products, proposing a temporary split, checking analytics) are examples of good handling, not a mandatory checklist. If the manager handled the objection convincingly in another way and the partner agreed — this is full compliance.
SCORING SCALE:
3 = Full compliance. The objection was handled substantively: a meaningful reply was given, a solution/compromise/alternative was proposed, the next steps were recorded or an agreement was reached. Also assign 3 if there was no objection in the conversation (the criterion is not applicable).
2 = Partial compliance. The objection was handled incompletely: a reply was given but was superficial; the solution or alternative was stated vaguely; the next steps were not clearly recorded. The partner was not left entirely without a reply, but the problem was addressed weakly.
1 = Non-compliance. The objection was effectively not handled: there is no substantive reply and no action plan, no solution was proposed, no next steps were agreed, the partner remained dissatisfied.
If the conversation is empty (no messages at all) — the criterion is not applicable, assign 3.
Briefly explain why this score was assigned, referring to specific fragments of the conversation.

Driving the Conversation to a Result

Evaluation window: default

Evaluates whether the manager drives the communication to a result or at least to a clearly agreed next step, without leaving the conversation hanging and without a status. It takes into account whether the manager responds to the partner's messages, keeps control of the process during prolonged silence and warns about delays. The criterion evaluates not the style of communication but specifically whether the issue was brought to a clear outcome. It helps track whether partner requests are being lost without completion.

Show prompt

You are evaluating the quality of a manager's work in a conversation with a partner against the criterion "Driving the Conversation to a Result". Evaluate the manager's actions (as well as those of other team members conducting the conversation on behalf of the company). Rely only on facts from the conversation, message history, message dates and other metadata. Do not make assumptions. Evaluate not the style of communication but whether the issue was brought to a result or a clear next step.
TASK: Determine whether the manager drove the communication to one of two outcomes: (1) the question/request was resolved, or (2) a clear next step or status was agreed. The conversation must not be left hanging without a result and without agreements.
WHAT COUNTS AS DRIVING TO A RESULT:
- the partner's question was resolved;
- or a clear next step / status was stated ("I'll do it by Monday", "passed it to the department, I'll come back with an answer");
- if the issue cannot be closed immediately — the manager showed that work is continuing and stated a next step or a new deadline;
- if the manager delivered a result (a reconciliation, status, instruction) and the partner confirmed it with a closing reply ("ok", "approved", "got it", "thanks, it works") and no further action from the manager is required — this counts as completing the conversation, even without a separate final message from the manager;
- if the partner took the next step upon themselves ("I'll check / come back with an update") — this is an agreed next step, and the absence of a follow-up from the manager is not a violation until a reasonable amount of working time for a check-in has passed (including when a weekend follows).
WHAT COUNTS AS A VIOLATION:
- the request was not driven to a result and no next step was agreed — the conversation was cut off in uncertainty;
- a message or request from the partner was left without a reply for 2 or more working days, when a reply was expected by context;
- inaction: the manager did not take the necessary steps (including a previously promised action) to continue or complete the process;
- in the event of a delay, the manager did not notify the partner and did not state a new deadline or next step.
IMPORTANT CLARIFICATIONS (protection against false positives):
- a closing reply from the partner ("ok / thanks / approved") which is not a question or request does not require a reply from the manager and is not treated as a "missed message";
- the mere fact that "the partner is not responding" is not an error; the error is the manager's inaction during prolonged silence — the absence of a check-in, status or a previously stated deadline after a reasonable amount of working time has passed;
- control of the process (a check-in, status, deadline) carried out by any team member on behalf of the company counts — do not attribute the error to personal authorship;
- if a key request from the partner arrived on a weekend or outside working hours, do not record a violation until a reasonable amount of working time for a reaction has passed within the window; record a violation only if the request still remained without a reply/status after working hours began;
- if the manager warned about a delay in advance and gave a new deadline — this is not a violation.
SCORING SCALE:
3 = Full compliance. The conversation was driven to a result or to a clearly agreed next step; the partner was not left without a reply and without a status; in the event of a delay the manager notified and stated a new deadline; it is clear from the conversation how the discussion ended or what will happen next.
2 = Partial compliance. The conversation is generally being conducted, but the outcome is vague: the next step is stated unclearly, the status was given late or incompletely, there is minor incompleteness. The partner is not abandoned, but the result is recorded weakly.
1 = Non-compliance. The conversation was left hanging: the request was not resolved and no next step was agreed; or the partner's message was left without a reply for 2+ working days; or the manager did not perform the necessary/promised action; or in the event of a delay they did not notify and did not give a new deadline.
If the conversation is empty (no messages at all) — the criterion is not applicable, assign 3.
Briefly explain why this score was assigned, referring to specific fragments of the conversation and dates.

Request Handling

Evaluation window: default

Evaluates how well the manager handles the partner's request: whether they understood its essence, showed that the request was taken into work, gave a substantive reply and recorded the agreements (including a post-meeting summary after calls). It helps track whether partners receive complete and documented answers to their requests.

Show prompt

You are evaluating the quality of a manager's work in a conversation with a partner against the criterion "Request Handling". Evaluate the manager's reaction (as well as that of other team members replying on behalf of the company). Rely only on facts from the conversation. Do not make assumptions.
TASK: Determine how well the manager handled the partner's request.
APPLICABILITY:
Evaluate only the reaction to explicit requests, questions or tasks written by the partner. The criterion is not applicable (assign 3) if:
- there is not a single partner request in the conversation;
- the partner's message is merely an acknowledgement or reaction ("ok", "thanks", "approved") that does not require a reply;
- the request is clearly addressed to another department/person (finance, accounting, a specific @handle) and does not require action or escalation from the manager;
- the question is contained in the manager's own message (a manager's question is not a partner request).
WHAT TO LOOK AT:
1. Taking into work: the manager showed that the request was received and understood. Do not require a formal "request received" acknowledgement if the manager moved straight to solving it or asked clarifying questions. If the solution depends on other departments/the tech team or on the partner's own readiness, the manager is not obliged to give a final "yes/no" or an exact date — it is enough to confirm that the request was received, state the next step and/or that an update is awaited, and give an expected timeframe for the update ("I'll check and come back today", "we're waiting until the weekend").
2. Completeness of the reply: the manager's reply closes the partner's question — the partner understands what to do next. Do not require a mandatory "yes/no" format: if the manager gave a sufficient explanation or confirmation in substance, the question is closed. Treat a reply as incomplete only if, after it, ambiguity remains that prevents the partner from acting (missing key data, terms or a next step) and the manager did not clarify it. Do not treat the absence of a final decision as an incomplete reply where the decision objectively does not depend on the manager or depends on the partner themselves, provided the manager stated that the request was taken into work / a next step / that an update is awaited.
3. Recording agreements: the agreements reached are clear from the conversation. In an ordinary text chat it is enough for the terms to be clearly stated along the way (agreed GEOs, rates, caps) — a separate summary message is not required. A post-meeting summary (a separate message with the outcomes) is mandatory after voice or video calls: if it is clear from the conversation that a call took place or was agreed, but there is no summary/outcome in the chat afterwards, this counts as a missing post-meeting summary.
SCORING SCALE:
3 = Full compliance. The request was understood and taken into work, a substantive reply was given that closes the question (or the wait for an update was correctly stated if the decision does not depend on the manager), agreements were recorded, and there is a post-meeting summary after a call.
2 = Partial compliance. The request was handled but with shortcomings: the reply is incomplete and leaves ambiguity, or the agreements were only partially recorded, or the post-meeting summary after a call is missing. The request was not ignored, but it was not handled to a fully satisfactory standard.
1 = Non-compliance. A partner request addressed to the manager, or clearly requiring their action, was ignored (there is no reaction and no acknowledgement of it being taken into work), or the recording of agreements / the post-meeting summary after a discussion is entirely absent.
If the criterion is not applicable (no partner request, only acknowledgements, a request to another department) or the conversation is empty — assign 3.
Briefly explain why this score was assigned, referring to specific fragments of the conversation.

Quality of Communication with the Partner

Evaluation window: default

Evaluates how correctly, politely and clearly the manager communicates with the partner: whether they observe basic standards of business courtesy, avoid rudeness or toxicity, convey information clearly and do not distort information about the company and its products. It takes the industry into account: an informal style is acceptable (professional slang, emojis, light informality) — the score is lowered only for outright rudeness or distortion of information.

Show prompt

You are evaluating the quality of a manager's work in a conversation with a partner against the criterion "Quality of Communication with the Partner". Evaluate only the manager's messages. Rely only on facts from the conversation. Do not make assumptions.
TASK: Determine how correctly, politely and clearly the manager communicates with the partner.
WHAT TO LOOK AT:
1. Tone and politeness: the manager observes basic standards of business courtesy, does not allow rudeness, harshness, toxicity or abusive language. Take the industry specifics into account: an informal style is acceptable — professional slang, emojis, diminutive forms, minor typos. Lower the score only for outright rudeness, toxicity or clear disrespect, not for informality as such.
2. Clarity of communication: the manager conveys information clearly — the partner understands what is meant and what to do next. This includes whether the manager explains the required actions to the partner in substance (not about speed, but about the essence of the instruction).
3. Accuracy of information about the company and its products: when the partner asks about the company, its role or the differences between products/brands (or when there is clear confusion), the manager explains this correctly, without distortion. Evaluate this only if the topic arose in the conversation.
SCORING SCALE:
3 = Full compliance. Communication is polite and correct, information is conveyed clearly, and details about the company and its products (if the topic arose) are stated without distortion.
2 = Partial compliance. Communication is generally correct, but there are shortcomings: information is conveyed insufficiently clearly and leaves ambiguity, or the explanation about the company/products is incomplete or slightly inaccurate. There is no rudeness.
1 = Non-compliance. There is a clear violation: rude, harsh or toxic communication, an absence of basic business courtesy, or distortion of information about the company/products (an incorrect explanation of the company's role or confusion between brands).
If there is no substantive communication from the manager to evaluate (the conversation is empty or contains only acknowledgements) — the criterion is not applicable, assign 3.
Briefly explain why this score was assigned, referring to specific fragments of the conversation.

Accuracy of Deal Terms

Evaluation window: default

Evaluates whether the manager correctly documents and manages the terms of the deal: whether KPIs, caps, payment and other key parameters are stated correctly, whether the terms are fully written out when links are issued, and whether the manager misleads the partner with contradictory or erroneous wording. It covers both the moment of issuing the link (the full set of mandatory fields) and the accuracy of the terms throughout the discussion of the deal.

Show prompt

You are evaluating the quality of a manager's work in a conversation with a partner against the criterion "Accuracy of Deal Terms". Evaluate the manager's messages (as well as those of other team members issuing links on behalf of the company). Rely only on facts from the conversation. Do not make assumptions.
TASK: Determine whether the manager correctly documented the issuing of the link and the terms of the deal — whether all mandatory parameters are written out and whether the KPIs and terms are correct.
APPLICABILITY:
The criterion applies only to conversations where the manager issues the partner a traffic link or agrees the terms of a deal. If there is no issuing of links and no agreement of terms in the conversation — the criterion is not applicable, assign 3.
WHAT TO LOOK AT:
- when issuing a link for traffic, all mandatory deal-term fields are written out in accordance with the team's rules (the set of fields is defined in the criterion settings — for example: source, project, GEO, deal terms, cap, postclick, KPI, links with landing page names, as well as other conditions important for running traffic);
- different traffic types may require different sets of fields (for example, a separate format for SEO) — check against the relevant set;
- KPIs are stated correctly and with an understanding of the consequences if they are not met;
- the deal terms are not contradictory and leave no ambiguity for the partner;
- the link is not issued "bare", without the accompanying mandatory terms.
SCORING SCALE:
3 = Full compliance. The link was issued with all mandatory fields, the KPIs and terms are stated correctly and in full, and the partner has all the parameters needed to work.
2 = Partial compliance. The link was issued, but the documentation is incomplete: some mandatory fields are missing or stated unclearly, the KPIs/terms are only partially written out. There is no critical error in the terms.
1 = Non-compliance. The link was issued incorrectly: the mandatory deal terms were not written out, the KPIs or terms were stated incorrectly, or the link was issued without any accompanying terms at all.
If the criterion is not applicable (no links are issued in the conversation) or the conversation is empty — assign 3.
Briefly explain why this score was assigned, referring to specific fragments of the conversation.

Reconciliation with the Partner

Evaluation window: default

Evaluates whether the manager sends the partner the reconciliation (the settlement for the period worked) on time, does so in the official partner chat and without errors in the data. Each team sets the sending deadline according to its own rules (for example, by a certain day of the month; if it falls on a weekend — by the next working day). The criterion applies only to conversations where a reconciliation is required and its deadline has arrived. It helps track discipline and accuracy in financial settlements with partners.

Show prompt

You are evaluating the quality of a manager's work in a conversation with a partner against the criterion "Reconciliation with the Partner". Evaluate the manager's actions. Rely only on facts from the conversation, message dates and available metadata. Do not make assumptions.
TASK: Determine whether the manager sent the reconciliation to the partner on time, in the official partner chat and without errors in the data.
REQUIREMENTS:
- the reconciliation was sent by the established deadline (by default — by the 10th of the month; if the deadline falls on a weekend/holiday — by the next working day);
- the reconciliation was sent in the official partner chat, not outside it;
- the reconciliation data is correct, without errors or distortions.
For traffic types where a reconciliation is not required (for example, SEO), the criterion is not applicable — assign 3.
DEADLINE RULE (protection against false positives):
You may only penalise for a missed deadline or a non-delivered reconciliation if it is clear from the conversation that the deadline has already arrived and the reconciliation was not sent or was sent late.
If the deadline has not yet arrived within the available conversation — do not penalise for "not sent", even if the reconciliation is still partial and it has been promised to be completed later; assign 3.
If it is impossible to determine from the conversation whether the reconciliation was sent on time (the conversation does not cover the deadline period, started after it, or ends earlier) — assign 3.
If the conversation covers several months — evaluate by the worst month.
RULE ABOUT THE PARTNER'S CONFIRMATION:
A message from the partner such as "reconciliation approved / approve / ok on the reconciliation" means that the partner has already received the reconciliation and agrees with it. This is NOT a request to "send the reconciliation" and NOT a sign that the reconciliation is missing — no violation is recorded on the basis of such messages.
SCORING SCALE:
3 = Full compliance. The reconciliation was sent on time, in the official partner chat, and the data is correct. Also assign 3 if the criterion is not applicable (a reconciliation is not required for this traffic type; the deadline has not arrived; the conversation does not allow the fact of sending to be determined).
2 = Partial compliance. The reconciliation was sent on time and in the right chat, but there is a minor inaccuracy in the data which was noticed/corrected and did not lead to a serious error in the calculations.
1 = Non-compliance. There is a clear violation: the reconciliation was not sent by the deadline that had arrived, or was sent late; it was sent outside the official partner chat; or it contains incorrect data.
If the conversation is empty (no messages at all) — assign 3.
Briefly explain why this score was assigned, referring to specific fragments of the conversation and dates.

Providing Analytics and Reporting

Evaluation window: default

Evaluates how regularly, clearly and usefully the manager provides the partner with analytics and reporting on results. It takes into account not only the fact of sending data, but whether the manager helps interpret the figures: explains the metrics, highlights problems and deviations, and draws conclusions and recommendations for improvement. Substantive analytics that help the partner make a decision are valued, rather than a formal forwarding of numbers. The criterion applies when reporting on results is appropriate in the conversation or is requested.

Show prompt

You are evaluating the quality of a manager's work in a conversation with a partner against the criterion "Providing Analytics and Reporting". Evaluate the manager's messages. Rely only on facts from the conversation and available metadata. Do not make assumptions.
TASK: Determine how regularly, clearly and usefully the manager provides the partner with analytics and reporting on results, and whether they help interpret the data rather than simply forwarding numbers.
APPLICABILITY:
Evaluate fully if the conversation contains a discussion of results, figures, payouts, conversions or traffic quality, or if the partner requests data/a report/an explanation of results. If there is no occasion for reporting in the conversation (the topic of results does not come up) — the criterion is not applicable, assign 3, and do not lower the score for an absence of analytics where it was not required.
WHAT TO LOOK AT:
1. Regularity and timeliness: the manager provides data when it is appropriate or expected; does not make the partner request the same thing repeatedly; meets the promised deadlines for reports, if any were stated.
2. Completeness and usefulness: the reporting contains data that is genuinely significant for the partner (conversions, volumes, payouts, traffic problems, dynamics by offer/GEO/source, reasons for drops and deviations). The reporting is substantive rather than formal.
3. Clarity of presentation: from the message the partner understands what happened, what the current results are, where there is a problem or deviation, and what to pay attention to. The data is not delivered "raw" and without explanation.
4. Interpretation: the manager does not limit themselves to forwarding figures — they explain what the metrics mean, highlight important changes, point out a problem, help understand the reason for a drop or a strong result.
5. Recommendations: the manager provides useful conclusions based on the data — what can be changed, what next step to take, which approach to test, how to improve the result.
SCORING PRINCIPLE:
Do not assign a high score merely for the fact of sending figures. Place greater value on clarity, interpretation and the practical usefulness of the analytics — whether it helps the partner understand the situation and make a decision.
SCORING SCALE:
3 = Full compliance. The manager provides substantive and clear analytics in a timely manner, helps interpret the metrics, highlights important conclusions and problems, and gives sensible recommendations or a next step based on the data.
2 = Partial compliance. Analytics is generally provided, but with shortcomings: the data is not always conveyed clearly, completely or on time; the interpretation is weak or inconsistent; the recommendations are generic, incomplete or weakly tied to the data.
1 = Non-compliance. The manager does not provide the required analytics where it was clearly needed; the reporting is late, incomplete or unclear; the partner receives figures without explanation and without the ability to understand what they mean; conclusions and recommendations are absent where they are obviously needed by context.
If the criterion is not applicable (there is no occasion for reporting in the conversation) or the conversation is empty — assign 3.
Briefly explain why this score was assigned, referring to specific fragments of the conversation.

Partner Reactivation

Evaluation window: 30 days

Evaluates whether the manager notices a decline in the partner's activity and tries to bring them back into work: whether they reach out themselves, identify the real reason for the decline and propose specific, situation-appropriate reactivation options. What is valued is not a formal check-in but a considered approach: understanding the reason for the drop and proposing a solution tailored to it. The criterion applies only where there are clear signs of a decline in activity or a pause on the partner's side. A successful outcome (whether the partner returned) is not required — the quality of the attempt is what is evaluated.

Show prompt

You are evaluating the quality of a manager's work in a conversation with a partner against the criterion "Partner Reactivation". Evaluate the manager's actions. Rely only on facts from the conversation and available metadata. Do not make assumptions.
TASK: Determine whether the manager notices a decline in the partner's activity and makes a considered attempt to bring them back into work.
APPLICABILITY:
Evaluate fully only if the conversation or metadata contains signs of a decline: the partner has stopped sending traffic, reduced volume, gone on pause or dropped out of active work. If there are no clear signs of a decline — the criterion is not applicable, assign 3, and do not lower the score for an absence of reactivation where there was no occasion for it.
WHAT TO LOOK AT:
1. Initiating contact: the manager reaches out to the inactive partner themselves, rather than passively waiting for them to write first.
2. Identifying the reason for the decline: the manager tries to understand why the partner reduced activity — clarifies the reason for the pause, asks about conversion, rates, offers, traffic, seasonality, priorities, technical/product/commercial obstacles. What matters is not simply "pinging" the partner, but getting to the real reason.
3. Proposing solutions: the manager proposes specific steps tailored to the situation — new offers, improved terms/rates, test budgets, help with creatives, new GEOs, alternative products, an analysis of the drop, a clear next step for a relaunch.
4. Relevance of the approach: the solutions are appropriate and tailored to the partner's context, rather than templated check-ins with no value or random proposals unrelated to the reason for the decline.
5. Moving towards a result: the manager moves the situation forward — agrees a next step, proposes a concrete plan. A successful return is not mandatory (it does not always depend on the manager), but there must be a genuine, substantive attempt.
SCORING PRINCIPLE:
Do not assign a high score merely for the fact of sending the partner a message. Place greater value on the attempt to understand the reason for the decline and to propose a solution appropriate to it. Do not lower the score simply because the partner did not reply, if the manager initiated the reactivation properly.
SCORING SCALE:
3 = Full compliance. The manager reached out to the inactive partner themselves, tried to understand the real reason for the decline, proposed specific and appropriate reactivation options, acted thoughtfully and moved the conversation towards a next step or a return to work.
2 = Partial compliance. There is an attempt at reactivation, but an average one: contact was initiated, but the reason for the decline was explored weakly; the proposals for a return are generic, imprecise or unconvincing; depth and strategy are lacking.
1 = Non-compliance. The manager does not attempt to reactivate the partner where there is a clear occasion to do so; limits themselves to a formal message without any attempt to understand the reason; proposes no solutions; makes no substantive attempt to bring the partner back into work.
If the criterion is not applicable (there are no signs of a decline in activity) or the conversation is empty — assign 3.
Briefly explain why this score was assigned, referring to specific fragments of the conversation.

Data extraction from communication

What request did the partner come with?

Type: CategoricalEvaluation window: default

Identifies the topic of the partner's enquiry — what request or question they came with. The attribute helps build statistics on the structure of enquiries: which requests come in most often and where the load on managers is concentrated.

Show prompt

Identify what request the partner came with in this conversation.
RULES FOR FORMING THE CATEGORY:
- use a short, generalised name for the request (2–5 words) that reflects the essence of the enquiry;
- keep the wording consistent: if a category matching the meaning has already occurred before, reuse its exact name rather than creating a synonymous variant (for example, do not create both "rate review" and "question about the rate" — these are one category);
- do not split requests that are close in meaning into different categories; generalise to a stable topic;
- the category must describe the TYPE of request, not retell the specific conversation.
IF THERE IS NO REQUEST:
If the partner did not come with a request (the conversation is empty, contains only acknowledgements/reactions such as "ok", "thanks", or the manager was the initiator) — assign the category "No request from the partner".
EXAMPLES OF TYPICAL CATEGORIES (a guide to the level of generalisation, not a fixed list):
- No request from the partner
- Request for offer/launch terms (rates, KPIs, bonuses, products, GEOs)
- Rate review
- Reconciliations
- Materials / links for placement (logo, creatives, replacing a tracking link)
Briefly explain, on the basis of which partner messages the category was chosen.

Was there an offer for … ?

Type: CategoricalEvaluation window: set manually

Checks whether a specific offer, GEO or product was proposed to the partner within a defined period. Helps track whether priority pushes are reaching partners. Categories: Yes / No / Not applicable. What exactly is checked (which GEO, offer or product) and for what period is defined per task.

Show prompt

Check whether the offer / GEO / product [INSERT NAME] was proposed to the partner during the defined check period [INSERT PERIOD]. Rely only on facts from the conversation and message dates. Do not make assumptions.
WHAT COUNTS AS A PUSH ("Yes"):
Count a push if the conversation (from either side — both in the manager's and in the partner's messages) contains a mention of the name of the GEO / offer / product being checked, or its spelling variants [LIST VARIANTS, for example: Norway, NO, Norge]. The push may take various forms — a direct proposal, an update on the GEO, a mention in the context of "it's already available for testing", a list of available GEOs containing this name, and so on.
WHAT DOES NOT COUNT AS A PUSH:
Do not count as a push a general question without a mention of the name being checked — for example, "which GEOs are ready for testing?" without a specific mention of the required GEO. A push exists only when the name being checked was actually stated.
CATEGORIES (choose strictly one):
- "Yes" — the offer / GEO / product being checked was mentioned within the check period.
- "No" — the check period is fully covered by the visible conversation, but there is no mention of the offer / GEO / product being checked.
- "Not applicable" — the period of visible messages does not fully cover the check period (there is insufficient data to judge).
Briefly explain, on the basis of which message and from which date the category was chosen.

Reason for the partner's decline / departure

Type: CategoricalEvaluation window: default

Identifies the reason why the partner reduced activity, went on pause or refused to continue working. Helps build statistics on churn drivers: what actually causes partners to be lost — conversion, rates, competitors, technical issues, and so on.

Show prompt

Identify the reason why the partner reduced activity, went on pause or refused to continue working in this conversation. Rely only on facts from the conversation — what the partner states as the reason themselves, or what clearly follows from the context. Do not make assumptions.
RULES FOR FORMING THE CATEGORY:
- use a short, generalised name for the reason (2–5 words) that reflects its essence;
- keep the wording consistent: if a reason matching the meaning has already occurred before, reuse its exact name rather than creating a synonymous variant (for example, do not create both "low conversion" and "poor conversion rate" — these are one reason);
- do not split reasons that are close in meaning into different categories; generalise to a stable topic;
- the category must describe the TYPE of reason, not retell the specific conversation.
IF THERE ARE NO SIGNS OF A DECLINE:
If the conversation contains no signs of reduced activity, a pause or a refusal from the partner (the partner is active, working as usual, the topic of a decline does not arise) — assign the category "No signs of a decline".
If a decline is visible but the reason is not clear from the conversation and the partner does not state it — assign the category "Reason not stated".
EXAMPLES OF TYPICAL CATEGORIES (a guide to the level of generalisation, not a fixed list):
- No signs of a decline
- Reason not stated
- Low conversion
- Unsatisfied with rates / payout terms
- Moved to a competitor / better terms elsewhere
- No suitable GEOs / offers
- Technical problems (tracking, postbacks, payments)
- Traffic problems on the partner's side (drop, quality, ad accounts)
- Change of priorities / focus on other projects
- Delays or problems with payouts
Briefly explain, on the basis of which messages the category was chosen.

Traffic type

Type: CategoricalEvaluation window: default

Identifies the type of traffic the partner works with (SEO, PPC, ASO, social media, etc.). Helps segment the partner base by traffic source: to understand which channels prevail and how the team's work is distributed.

Show prompt

Identify the type of traffic the partner works with in this conversation. Rely only on facts from the conversation — what the partner or the manager states directly about the channel/traffic source, or what clearly follows from the context (mentions of keywords, domains, campaigns, the nature of the work). Do not make assumptions.
RULES FOR FORMING THE CATEGORY:
- use the standard short name for the traffic type;
- keep the wording consistent: reduce mentions to a single name rather than creating synonyms (for example, "context", "contextual advertising", "Google Ads" all belong to one category, "PPC"; "seo", "organic", "search traffic" — one category, "SEO");
- do not split one traffic type into different categories because of differing wording by the partner.
SELECTION MODE:
If the partner works with several traffic types at once — list them all. If there is one type — list one.
IF THE TYPE IS NOT DETERMINED:
If the conversation contains no information about the partner's traffic type (the topic does not come up, there are no indications) — assign the category "Traffic type not stated".
EXAMPLES OF TYPICAL CATEGORIES (a standard set, for guidance):
- Traffic type not stated
- SEO (organic search)
- PPC (contextual / search advertising)
- ASO (app store optimisation)
- Facebook / social media
- Influencers / bloggers
- Push / popunder
- UAC / media buying
- Email / newsletters
- Streams / Twitch / YouTube
Briefly explain, on the basis of which mentions the category was chosen.

Outcome of rate/payout negotiations

Type: CategoricalEvaluation window: default

Identifies how negotiations over the rate or payout terms ended: whether the partner accepted the terms, whether an improvement was agreed, whether they were refused and how they reacted. Helps build statistics on negotiation dynamics: how often terms have to be improved for the sake of retention, how often partners accept the initial proposal, and in how many cases a refusal leads to dissatisfaction or departure.

Show prompt

Identify how negotiations over the rate, payout or commercial terms with the partner ended in this conversation. Rely only on facts from the conversation. Do not make assumptions. Choose strictly one category.
WHAT COUNTS AS RATE/PAYOUT NEGOTIATIONS:
A discussion in which the partner asks to change the rate, payout amount, payment terms, hold, minimum FTD or other commercial terms — or bargains over them. Simply informing the partner about the terms, without a request to change them, does not count as negotiations.
CATEGORIES (choose strictly one):
- "No rate bargaining took place" — the conversation contains no discussion of changing the rate/payout/terms (the attribute is not applicable to this conversation).
- "Partner accepted the proposed terms" — the partner agreed to the proposed terms without changes in their favour.
- "Terms improved in the partner's favour" — as a result of the negotiations, the rate/payout/terms were changed in the partner's favour, and they accepted this.
- "A compromise was reached" — the parties settled on an intermediate option: a partial improvement, or an improvement subject to counter-conditions (for example, a higher rate upon meeting KPIs).
- "Improvement refused, partner accepted" — the manager did not improve the terms, but the partner accepted this and continued working.
- "Improvement refused, partner dissatisfied" — the manager did not improve the terms, the partner remained dissatisfied, refused or went on pause.
- "The question was left unresolved" — negotiations began but did not conclude: no decision was made, and the outcome is unclear from the conversation.
RULE FOR AMBIGUOUS CASES:
If the outcome of the negotiations cannot be determined from the conversation (there is no final reaction from either side) — choose "The question was left unresolved" rather than guessing the result.
Briefly explain, on the basis of which messages the category was chosen.

Was there an escalation?

Type: CategoricalEvaluation window: default

Identifies whether the conversation required an escalation — the involvement of a team lead, another department or a senior employee to resolve the partner's issue. Helps build statistics on case complexity and the load on senior staff: how often managers cannot cope on their own, which topics most often go up the chain, and where the bottlenecks in processes are.

Show prompt

Identify whether there was an escalation in this conversation — the involvement of a team lead, a manager, another department or a senior employee to resolve the partner's issue. Rely only on facts from the conversation. Do not make assumptions. Choose strictly one category.
WHAT COUNTS AS AN ESCALATION ("Yes"):
A situation where the partner's issue goes beyond what the manager resolves on their own, and someone else genuinely becomes involved in resolving it. Indicators:
- the manager explicitly passes the issue to a team lead / manager / another department (finance, tech support, product, media buying) for resolution;
- a senior employee joins the conversation and takes part in resolving the issue;
- the manager tells the partner that the issue has been passed up / referred to those responsible, and that their decision is awaited.
WHAT DOES NOT COUNT AS AN ESCALATION:
- an ordinary mention of colleagues without passing the issue on for resolution ("I'll check with the team and come back" as part of the manager's routine work — this is not an escalation if the manager continues to handle the issue themselves);
- generic phrases such as "I'll pass this on to the developers / share the feedback with the team" as a reaction/reassurance, without genuinely transferring the issue to someone else for resolution;
- situations where the manager copes on their own and merely consults in passing.
CATEGORIES (choose strictly one):
- "Yes" — there was an escalation: the issue was genuinely passed on, or a senior employee / another department became involved in resolving it.
- "No" — there was no escalation: the manager handled and resolved the issue themselves, or the conversation contained no issue requiring resolution at all.
Briefly explain, on the basis of which messages the category was chosen.

Topic of the complaint/problem

Type: CategoricalEvaluation window: default

Identifies the topic of the partner's complaint or problem — what exactly the negativity, grievance or failure concerns. The attribute is only triggered when there is dissatisfaction or a problem. Helps build statistics on the drivers of dissatisfaction: what partners complain about most often, and where the pain points in the terms or communication lie.

Show prompt

Identify the topic of the partner's complaint or problem in this conversation — what exactly the dissatisfaction, grievance or failure concerns. Rely only on facts from the conversation. Do not make assumptions.
APPLICABILITY:
The attribute is only triggered when there is a complaint, dissatisfaction, grievance or problem on the partner's side. An ordinary neutral request or question without negativity does not count as the topic of a complaint.
RULES FOR FORMING THE CATEGORY:
- use a short, generalised name for the topic (2–5 words) that reflects the essence of the problem;
- keep the wording consistent: if a topic matching the meaning has already occurred before, reuse its exact name rather than creating a synonymous variant (for example, do not create both "low conversion" and "poor conversion rate" — these are one topic);
- do not split problems that are close in meaning into different categories; generalise to a stable topic;
- the category must describe the TYPE of problem, not retell the specific conversation.
IF THERE IS NO COMPLAINT:
If the conversation contains no complaint, dissatisfaction or problem (the partner is neutral, comes with an ordinary request, or the conversation is positive) — assign the category "No complaint / problem".
EXAMPLES OF TYPICAL CATEGORIES (a guide to the level of generalisation, not a fixed list):
- No complaint / problem
- Low conversion
- Dissatisfaction with rates / payout terms
- Delay or problem with a payout
- Technical problems (tracking, postbacks, links)
- Slow response / poor communication
- Problem with an offer (closed, restricted, no slots)
- Incorrect data in the reconciliation
- Disagreement with the terms of cooperation
Briefly explain, on the basis of which messages the category was chosen.

Number of messages

Type: NumericEvaluation window: default

Counts how many messages the manager sent in the chat during the evaluation period. The attribute helps measure managers' activity in quantitative terms: to see the workload and compare engagement across managers. It is used for statistics and analysis.

Show prompt

Count the number of messages sent by the manager in this conversation during the evaluation period. Count only the manager's messages — do not count messages from the partner, from any other company representative, or automated/system messages. Return a number.

Number of initiated follow-up messages

Type: NumericEvaluation window: default

Counts how many quality follow-ups the manager initiated themselves during the evaluation period in order to move the discussion forward: reminders about agreements, status clarifications, requests for missing data, proposed solutions, or bringing the partner back to action. Only meaningful follow-ups are counted, not empty check-ins and spam. The attribute helps measure the manager's proactivity — how far they drive conversations themselves rather than waiting for the partner. Unlike the total number of messages, what matters here is the quality and initiative of repeated touchpoints.

Show prompt

Count the number of quality follow-up messages that the manager initiated themselves in this conversation during the evaluation period. Count only the manager's messages. Return a number.
WHAT COUNTS AS A QUALITY FOLLOW-UP:
A meaningful repeat message from the manager that returns to an already open topic and moves the conversation forward. Such a message contains new value for the partner, for example:
- a useful reminder about an agreement;
- a status clarification on an open question;
- a request for missing data needed to make progress;
- a proposed solution or next step;
- bringing the partner back to active steps on a previously discussed topic.
WHAT DOES NOT COUNT AS A FOLLOW-UP (do not include in the count):
- the first initiating message in the conversation (this is not a repeat touchpoint);
- empty check-ins and formal reminders with no value ("?", "any news?" with no substance);
- spam and repetition of the same thing without new information;
- ordinary replies from the manager to a message from the partner (a follow-up is the manager's initiative, not a reaction to the partner's message).
Return the final number of quality follow-up messages.

Response time

Type: NumericEvaluation window: default

Measures how long on average it takes the manager to reply to a partner's request. The attribute helps build statistics on reaction speed: to compare managers, track dynamics and identify periods with slow responses.

Show prompt

Determine how long it took the manager to reply to the partner's requests in this conversation. Rely only on the dates and times of the messages and the roles/labels of the authors. Do not make assumptions. Return a number (the average response time in minutes).
WHAT COUNTS AS A PARTNER REQUEST (the starting point):
Measure response time only from messages in which the partner themselves initiates a request — asks a question, makes a request or seeks a clarification that requires a reaction from the manager. Before counting, check the message author by role/label.
DO NOT use as a starting point:
- messages from the manager themselves;
- the partner's replies to a manager's question/check-in ("ok", "in progress", "yes/no", a status) — these are not initiating requests;
- closing replies from the partner without a request ("thanks", "got it").
HOW TO COUNT:
- for each initiating partner request, find the manager's first substantive message in reply (an acknowledgement, a clarifying question, "checking now", a solution) and calculate the time difference;
- a "reply" is the manager's first reaction, not necessarily the final delivery of all materials;
- if there are several such requests — calculate the response time for each and return the average value.
IF THERE IS NO DATA:
If the conversation contains no initiating partner requests, or the request received no reply from the manager at all within the conversation — do not invent a value; return an empty value / 0 and note that there are no applicable requests.
Return the final number (the average response time in minutes).

Objection handling effectiveness

Type: NumericEvaluation window: default

Shows what percentage of the partner's objections and complaints were closed positively — that is, the partner accepted the explanation, continued working or improved the terms. It is calculated as the ratio of successfully handled objections to the total number of objections in the conversation. The attribute helps measure not the fact of handling but its effectiveness: how far managers genuinely win over dissatisfied partners.

Show prompt

Determine what percentage of the partner's objections and complaints in this conversation were closed positively. Rely only on facts from the conversation. Do not make assumptions. Return a number as a percentage.
WHAT COUNTS AS AN OBJECTION / COMPLAINT:
A clear signal of dissatisfaction, a grievance, refusal or disagreement from the partner. Examples: "poor conversion", "doesn't work for us", "too expensive", "too little", "I won't place it", "no slots", "competitors are better", bargaining over rates/terms/KPIs/deadlines.
DOES NOT count as an objection:
- a message about an action/result within a previously agreed plan, without dissatisfaction;
- informational questions without signs of a grievance;
- closing/neutral replies ("ok", "thanks").
WHAT COUNTS AS A SUCCESSFULLY CLOSED OBJECTION:
An objection is considered closed positively if, after the manager handled it, it is evident that:
- the partner accepted the manager's explanation or argument (agreed, withdrew the grievance);
- or the partner continued working / agreed to continue;
- or the terms were improved as a result and the partner accepted this.
DOES NOT count as successful:
- the objection was left without a reply or without a result;
- the partner remained dissatisfied, did not agree, or refused;
- the outcome is unclear from the conversation (there is no reaction from the partner after it was handled).
HOW TO COUNT:
1. Count the total number of objections/complaints from the partner in the conversation.
2. Count how many of them were closed positively (by the criteria above).
3. Return the percentage: (successfully closed / total number) × 100.
IF THERE ARE NO OBJECTIONS:
If there is not a single objection or complaint in the conversation — do not invent a value; note that there were no objections (the attribute is not applicable to this conversation).
Return the final number as a percentage.