Lost deal analysis: how to use lost deals to grow sales

Analysis of lost deals helps you understand where and why a company loses potential customers. These observations are then turned into improvements to the product, positioning, and sales process. It shows why deals are lost and uncovers the real reasons for customer rejections instead of the convenient versions from CRM comments. It also answers the question of how to increase sales conversion without extra discounts and heroic efforts.

This article will be useful for sales heads and managers, marketers, and business owners. There will be tables, formulas, and an action plan, with no lectures and no fluff.

Analysis of lost deals is a systematic review of opportunities that did not end in a sale. Its goal is to find the reasons for lost deals and recurring patterns. It’s important to distinguish between two levels here. Reviewing a single deal answers the question “what went wrong with this customer.” Regular analytics across the entire flow of losses answers a different question: why deals are lost overall and where exactly the system fails. The first is like analyzing a single chess game, the second is like statistics for the whole season.

In practice, analysis of the reasons for lost deals lets you identify recurring rejection scenarios. For example, if several customers with similar requests leave after receiving a commercial proposal, you should check not only the managers’ work but also the proposal itself: the terms of cooperation, timelines, and how clear the value is to the buyer.

Good analytics helps you make quite down-to-earth decisions. You adjust funnel stages, refine your target audience, revisit your offer, and rewrite scripts. At the same time, you stop spending budget on channels that bring in “tourists” instead of buyers. A simple scheme looks like this: data on the loss → identifying the reason → change → checking the result. If you skip the last step, you end up with a beautiful report that’s of no use.

The goal of the analysis is to improve the sales system. There’s no need to look for a manager to blame or to win back a departed customer at any cost. As soon as the review turns into a public flogging, the team starts choosing the safest explanations instead of the most accurate ones. That’s why you first need to agree on what counts as a loss at all.

Before analyzing anything, agree on statuses. Otherwise, the “lost” pile will include both a customer who chose a competitor, a duplicate request, and a lead with a wrong phone number. These are different stories, and counting them in one basket means spoiling your statistics. A good working rule: a deal is a separate purchase opportunity with a specific need, buyer, and subject of the offer. A request without qualification remains a lead, and losses at that stage should be counted separately. You can read more about why businesses lose leads in a separate article.

Pay special attention to the “postponed” and “customer went silent” statuses. Postponing a decision is justified only if the condition for returning and the date of the next check are known. For example, a customer from Lviv is waiting for next quarter’s budget to be approved. The phrase “let’s catch up sometime” doesn’t count as an active forecast. Silence doesn’t prove rejection either, so an honest entry reads “outcome unknown.” There’s no point in setting a universal auto-closing period. It will be different for an online store and a corporate implementation, so rely on your own deal cycle history.

Once the statuses are set, you can look at how your approach differs from the broader method.

Lost deal analysis focuses only on lost deals. It answers the question “why did we lose” and shows why deals fail in the first place. Win/loss analysis is broader because it compares won and lost deals, and sometimes adds a third group, no decision or postponed. This kind of win-loss analysis lets you see not only the reasons for loss but also what sets apart the deals you consistently win. It’s like studying flights only after crashes versus also studying how successful flights land.

International providers of such solutions, for example Clozd, build their practice on direct buyer feedback and include both wins and losses in the sample. But their claims of effectiveness should be treated as advertising rather than independent research. The method is useful, but it doesn’t guarantee conversion growth. To get started, analysis of failed deals is enough, and it’s better to add wins once your basic tracking is in order. Then the analysis of won and lost deals will show which factors really distinguish the outcomes.

What you can compare in win/loss analysis:

• involvement of the decision-maker
• speed of preparing the commercial proposal
• the customer’s segment or industry
• the specific product
• availability of a suitable case study
• quality of identifying customer needs at the discovery stage

If “too expensive” shows up in both won and lost deals, price is most likely not the main differentiator. For such comparisons to work, you need clean data, so next we’ll cover what to collect in your CRM.

The quality of your conclusions depends directly on how complete and consistent your data is. If one manager writes “expensive,” another “budget,” and a third leaves the field empty, analysis of lost deals turns into guesswork. You need a minimal card where some fields are filled in automatically, and the manager is left with what requires human understanding. That’s the content of the conversation, the context of the decision, and the next step. Order in your cards is directly tied to CRM data hygiene.

A separate note on customer comments: without them, analysis of customer rejections remains a formality. A “too expensive” tag explains nothing. Maybe the customer has no money. Maybe it’s hard for them to pay upfront. Or maybe they didn’t see a result that justifies the costs, or compared incomparable configurations. In such cases, handling the price objection comes in handy. A useful comment answers four questions: who communicated the decision, what exactly they said, when it happened, and what the conclusion is based on. Also save the last meaningful stage and the history of the amount. If a manager zeroes out the value before closing, the estimate of lost potential becomes meaningless. And if your CRM has several currencies (UAH, EUR, USD), agree on a conversion rule in advance, otherwise the exchange rate will start “improving” your sales.

At the same time, analysis of the reasons for customer rejections becomes more accurate when CRM comments are based on buyers’ actual words rather than managers’ assumptions. That’s why it’s important to record separately whether the customer confirmed the reason for rejection or the team determined it on its own based on the interaction history.

With the data collected, it’s time to sort the reasons onto shelves.

Start with a short reasons list, which in CRMs is often called lost reason or “лост ризон,” and don’t let it turn into a dumping ground. A good scheme stores three things separately: the outcome, the reason for loss, and the level of confirmation. “Chose a competitor” is an outcome, but not yet a reason. “Too expensive” is the buyer’s assessment, not the mechanism of rejection. And “the manager assumes” speaks to the level of evidence. When all of this sits in one list next to “couldn’t reach” and “duplicate,” there’s nothing to compare. For your report, choose one main reason for losing a deal and several contributing factors, and show the unknown as a separate “not established” category.

Distinguish the stated reason from the root cause. The objection “too expensive” can easily hide unclear value or an inconvenient payment format. In that case, restructuring the payment stages helps, while a discount only eats into your margin. The same goes for external circumstances. Don’t create a universal “situation in the country” category; it’s too broad to manage. Record an observable fact: funding suspended, agreed deadline missed, no implementation team, delivery unavailable. The reasons for failed deals are worth evaluating both by the number of deals and by lost revenue, because a rare but large problem can cost more than a frequent small one.

For the reasons for losing a deal to be determined objectively, it’s not enough to focus only on the buyer’s last objection. For example, if a customer named a high price, it’s important to find out which offers they compared your product with, which conditions turned out to be decisive, and who made the final purchase decision.

Once the reasons list is ready, we move on to the procedure itself.

It’s best to keep the sequence of actions simple, otherwise it will die in the second week. The main requirement for the sample: it must be sufficient and comparable by period. Reviewing only the most recent or most frustrating loss means analyzing emotions. Sometimes 20 closed deals are enough for first hypotheses, sometimes you need hundreds, and it all depends on the segment and cycle length. If you mix corporate implementations, retail orders, and real estate purchases, the same sample size won’t give the same depth.

The main rule after launching changes: check them in the next period. A rise in conversion next to the rollout of a new idea doesn’t prove the idea is the cause. Especially in Ukraine, where demand and the operational availability of customers can change quickly. If possible, compare parallel groups rather than “before and after.” For small companies with little data, it’s more honest to combine a pilot, qualitative evidence, and accumulating observations than to celebrate every short-term spike.

Step-by-step analysis scheme:

1. Define the period and scope of the sample, for example closed qualified deals for the quarter.
2. Check the data, find duplicates, empty reasons, and closures without a stage.
3. Assign categories to deals using the single reasons list.
4. Compare groups by key parameters: segment, source, amount, manager.
5. Find root causes through comments, correspondence, and, where possible, a conversation with the customer.
6. Formulate hypotheses and assign actions with owners.
7. Check the result in the next period and record the conclusion.

An example of a cycle. Observation: in the small online store segment, people often write “too expensive” after the demo. Hypothesis: customers don’t see payback in the first months. Action: add a payback calculation to the demo using a similar store from Kyiv as an example. Measurement: the share of transitions from demo to payment in this segment over the next six weeks. Now let’s look at which segments to search for such observations in.

Overall statistics like to hide contrasts. “We lose 60% of deals” sounds scary but doesn’t say where it hurts. Maybe everything is stable in one product and failing in another, or one lead source brings in nothing but noise. Start with the first level of segmentation: product, customer type, source of inquiry, opportunity size, and creation period. Then add attributes for a specific hypothesis, such as payment method or delivery type. Don’t cross all fields at once, otherwise you’ll get a pile of tiny groups where chance looks like a pattern.

For managers, it’s not enough to look at the loss rate. Compare portfolios: the quality of incoming leads, average deal size, segments, demand maturity. A strong manager with a “cold” flow may lose more than an average one with hot leads. Stage analysis shows bottlenecks (a funnel stage audit helps here), while a time slice helps you notice seasonality and the effects of process changes. For long cycles, use a cohort report, where you follow opportunities created in a certain period after the same observation time. A report by closings and a cohort report answer different questions.

Segments show where to look. To assess the scale of the problem, you need numbers.

A qualitative review should be backed by a few simple numbers. The basic indicators: loss rate, win conversion, losses by stage, average amount and potential revenue of lost opportunities, and the duration of the lost cycle. The loss rate formula is simple: number of lost deals ÷ number of all closed deals × 100%. In the denominator, include only closed deals for a comparable period. This is the basis for analysis of closed and lost deals. Analyze open opportunities separately, otherwise a company with a pile of stalled deals will look more successful than an honest team that closes quickly.

Example. In a quarter, 80 deals were closed, 28 won, 52 lost. Loss rate: 52 ÷ 80 × 100% = 65%. Win conversion: 28 ÷ 80 × 100% = 35%. If the average proposal amount is 40,000 UAH, the lost proposals totaled 52 × 40,000 = 2,080,000 UAH. But that’s the scale of the opportunities under consideration, not lost revenue. You can’t assume that all these customers would have bought if the problem hadn’t existed. There’s no universal benchmark here, so compare the result with the company’s history, sales type, and segment. For B2C, also check sales after cancellations and returns, otherwise growth in “order conversion” may turn out to be cosmetic. How to properly perform an assessment of funnel conversion is covered in detail separately.

At the same time, analysis of closed and lost deals shouldn’t be limited to a single conversion metric. Also compare the number of deals, their amounts, and the distribution of losses across funnel stages. That way you’ll see where the company loses more potential customers and at which stages it misses the largest contracts.

Metrics give you the picture, but that picture is easily distorted if the analysis is done with typical mistakes.

The most common mistake: taking “price” or “competitor” as the reason and stopping there. That’s the customer’s assessment or the name of a choice, not a mechanism. If you build decisions on it, the company will start handing out discounts when the problem is unclear value. The second mistake is that the CRM field is mandatory but the manager doesn’t know the reason. They’ll choose something “safe,” and a neat chart will scale a guess. The solution is simple: allow a “not established” category and store the level of confirmation.

Маленькая выборка. Three losses in a month is not a pattern. Conclusions drawn from them lead to sharp turns in scripts and offers that later have to be rolled back.

Смешение замороженных и окончательных проигрышей. A customer who will come back in a quarter ends up in “lost.” The forecast is understated, and nobody plans a repeat contact.

Выбор удобных данных. When a sales head checks only the deals that confirm their version, the analysis turns into self-affirmation. Selection rules are best fixed in advance.

Сравнение менеджеров без учета лидов. The same loss percentage can hide different problems. First compare within similar groups.

Анализ без владельцев изменений. A report without a responsible person and a check date turns into a slide that everyone praised and forgot.

If you avoid these traps, what remains is the main thing: turning conclusions into actions.

The unit of an analysis result should be not a chart but a verifiable initiative. It describes the problem, the affected segment, evidence, the assumed mechanism, the change, the owner, the metric, and the check date. A note like “customers don’t see value” is too vague. A working hypothesis sounds different, for example “small stores don’t understand payback after the demo, so we add a calculation to the presentation.” Until it’s tested, it’s only a hypothesis, even if it sounds convincing in a meeting.

Prioritize by three factors: frequency of the reason, potential revenue, and ability to influence it. For each change, assign an owner, a deadline, and a result metric. And don’t treat a discount or tighter control over managers as an automatic answer to every loss. If the problem is delivery times, objection-handling training won’t help. If the customer doesn’t see value, more frequent follow-ups will just increase the number of touches. A lost customer’s request is also not a reason to immediately build a new feature. Check whether it was a mandatory condition, whether it repeats in the target segment, and whether it breaks the product strategy. And watch for side effects: after refining your advertising, there may be fewer leads but more sales, so evaluate by target opportunities and completed sales.

What remains is to build all this into your everyday work rhythm.

The basic rules are these: a single list of loss reasons, mandatory card fields, clear closing criteria, and a regular review. Many CRMs technically separate the Won and Lost stages, but the rules for using them, including the reasons for closing a deal as a loss in Lost, the company must define itself. A good CRM collects and segments data for analysis of reasons for closing deals, but it doesn’t replace live comments, interviews, and interpretation. A mandatory field without the option to choose “not established” produces formal filling-in, not knowledge. Check how automations work not only in the manager’s interface but also during imports and bulk updates, otherwise a closure without data will slip past the rules.

To improve data reliability, analysis of reasons for closing deals is best started by checking the filled-in CRM cards. Pay attention to the cases where the reason was confirmed by the customer and where the manager indicated it as an assumption. This lets you separate real patterns from recording errors and make decisions based on reliable statistics.

As for rhythm, devote the weekly review to information quality. That means closures without grounds, unknown outcomes, overdue postponements, and lost stages. The monthly meeting is for recurring factors and the status of initiatives. Review the classifier less often, when the product, audience, or sales method changes, and introduce the new version with a date and mapping rules. Divide roles like this: the manager records facts, the sales head checks completeness, an analyst or sales operations is responsible for definitions, and marketing and product receive relevant conclusions. And don’t punish people for the growth in recorded losses after you’ve put things in order, otherwise the data will deteriorate again.

If you talk to customers for the analysis of customer rejections, remember data protection. The Law of Ukraine on Personal Data Protection requires a lawful purpose for processing and proportionality of data. So explain the purpose of the conversation, ask for consent to record, and don’t collect sensitive details if a business fact is enough. The specific scheme for transferring data to external services should be checked with a lawyer.

Launch checklist:

• agree on what a deal is and what counts as a win
• set up a reasons list with a “not established” category
• make the loss stage, date, amount, and a templated comment mandatory
• check how the CRM counts cancellations, returns, and costs
• schedule weekly and monthly reviews with owners
• prepare a template invitation for the customer to a short conversation

Lost deals are not just a minus in a report but a source of data on barriers in the product, marketing, and sales process. A useful analysis of lost deals relies on reliable data, a clear classification of reasons, comparison of comparable groups, and customer feedback. It doesn’t look for someone to blame and doesn’t promise miracles. It helps you make decisions that can be verified with numbers. Start small. Audit the cards of lost deals for the last quarter, pick one recurring scenario, and formulate a hypothesis for it. Then measure the result in the next period, not by individual stories. If after that you see that the reasons for customer rejections have become clearer and conversion has moved in the right direction, it means the system is working.