Finding the Real Reasons Behind Lost B2B Sales: A Case Study

Oct 16, 2026

Case Study

The most reliable way to find out why you're losing B2B deals is not to trust the "loss reason" field in your CRM. Instead, you should combine a structured loss reason taxonomy, independent win/loss interviews and cross-analysis of pipeline stage data. Labels like "price" or "timing" entered by sales reps usually reflect the easiest explanation rather than the real one. The actual cause is typically the wrong target audience, a weak discovery call, failure to reach the decision-maker or an inability to make the value tangible.

In this post, we'll walk through a representative case compiled from work we've carried out at Corner Group and anonymized to protect client confidentiality. We'll explain step by step how a mid-sized B2B SaaS company re-examined its lost opportunities, the steps it took to uncover the real loss reasons and how it fed those findings back into its sales process. The figures we share are illustrative values used to make the method concrete. Your numbers will differ depending on your industry, product and sales model.

Why doesn't lost sales analysis work at most companies?

Many B2B companies believe they're doing loss analysis. In reality, what they're usually doing is pulling a CRM report at the end of the quarter and concluding that "most losses are due to price." This approach has three fundamental problems:

The data is entered by the person who lost the deal. When a sales rep loses an opportunity, they tend to pick a reason that won't call their own performance into question. That's why "price" and "no budget" are always inflated.
The categories are too broad. A label like "competitor" or "timing" doesn't tell you which competitor you lost to, at what stage or for what reason. A category you can't act on isn't analysis.
Stage information is ignored. An opportunity lost during discovery and one lost after the proposal are not the result of the same problem. Lumping them together into a single "lost" pool blurs the picture.

Most of these problems actually come down to data discipline. We covered how to establish that discipline in multi-brand structures in detail in our case study on CRM data quality in a multi-brand holding company.

Case background: what problem was the company facing?

The company in this case was a B2B SaaS firm selling operations management software to mid-sized and large enterprises. The sales team consisted of SDRs and account executives. Lead sources were spread across outbound email, LinkedIn and web forms. The company used HubSpot as its CRM.

Here's the picture leadership was looking at:

Pipeline volume was growing steadily, but the win rate had been flat for several quarters.
According to the CRM report, more than half of all losses were tagged "price."
Based on this data, leadership was considering loosening its discount policy and pricing more aggressively.

We were asked to verify, before any discount decision was made, whether "the price problem was really a price problem." As a sample, we scoped in roughly 120 lost opportunities closed over the previous two quarters.

First step: what was behind the "price" label?

Our first move was to reconstruct the story of each opportunity without touching the loss labels in the CRM. To do this, we looked for answers to the following questions for every lost opportunity:

At what stage was the opportunity lost?
Was the decision-maker or budget holder involved in the process?
Did the discovery call cover the customer's business problem and the cost of that problem?
Was the proposal tied to the concrete outcomes the customer had described?
What alternative did the customer choose: a competitor, an in-house solution or doing nothing at all?

To answer these questions, we reviewed CRM notes, email threads, meeting recordings and proposal documents. A striking picture emerged within the very first week. In a significant share of the opportunities tagged "price," the decision-maker had never attended a meeting. The customer's problem had only been defined superficially. In other words, the customer found the price too high because they had never clearly seen the value being offered.

There's an important distinction here: "The price is too high" is not a reason; it's a symptom. The real question is why the customer didn't perceive enough value to justify that price.

How do you classify loss reasons correctly?

Since the existing categories weren't enough to act on, we created a new two-tier loss reason taxonomy. The first tier defined the main category, and the second tier defined the specific sub-reason.

Main categories

Fit-related losses: The company doesn't match the ideal customer profile, or the product doesn't meet the need.
Process-related losses: Insufficient discovery, the decision-maker wasn't reached, weak follow-up.
Value-related losses: The customer couldn't see the ROI, and the business impact wasn't made tangible.
Competition-related losses: A specific competitor was chosen for a specific reason.
No decision: The customer chose to stay with the status quo.
External factors: Reasons outside your control, such as an acquisition, a merger, a budget freeze or organizational changes.

Sub-reasons

We defined 3–5 concrete sub-reasons under each main category. For example, the "Process-related" category included:

The decision-maker was not involved in the process
The buying process and its steps were not clarified
Post-proposal follow-up was delayed by more than 7 days
The technical evaluation (demo/POC) was not customized to the customer's scenario

The critical rule of this structure was: Every sub-reason must translate into an action with a clear owner. The sub-reason "decision-maker not involved" pointed to the sales methodology, while "non-ICP company" pointed to the lead generation strategy.

How do you run win/loss interviews?

The CRM data and document review gave us a hypothesis, but we couldn't validate it without hearing the customer's voice. So we scheduled short win/loss interviews with a sample of lost opportunities.

Who should conduct the interviews?

The interviews were conducted not by the sales rep who managed the opportunity, but by an independent person. Customers find it hard to be honest with someone who has been trying to persuade them. A neutral interviewer can much more easily collect feedback like "Honestly, your rep never really understood us."

What questions should you ask?

In 20–25 minute interviews, we used the following set of questions:

What prompted you to look for this kind of solution? What was the most important problem you wanted to solve at the time?
Who was involved in the evaluation process, and who made the final decision?
What did you like most about our solution, and where did you hesitate?
What did you ultimately choose, and what were the two or three factors that drove that choice?
How decisive was price in your decision? If the price had been the same, would your decision have changed?
Was there anything you wanted to hear from us during the process but didn't?

The fifth question was especially valuable. A significant share of the answers to "If the price had been the same, would your decision have changed?" were "probably not." This clearly showed that price was a secondary rather than a primary factor.

How many interviews are enough?

Rather than chasing perfect statistical representativeness, a practical rule is to keep interviewing until patterns start repeating. In this case, after roughly 15–20 interviews, no new reasons were emerging and the main themes had become clear.

How did we rebuild the data structure in the CRM?

To keep the analysis from being a one-off project, we needed to embed the findings in the CRM permanently. We made the following changes in HubSpot. The same logic can be applied in Pipedrive and Zoho using custom fields and required field rules.

Two-tier loss reason field: Dependent dropdown lists for the main category and the sub-reason.
Required fields: When an opportunity was moved to the "Lost" stage, the record couldn't be saved without a main category, a sub-reason and a short explanation.
Competitor field: When "Competition" was selected, specifying which competitor the deal was lost to was mandatory.
Stage stamp: The stage at which the opportunity was lost was recorded automatically.
Decision-maker involvement: A field asking "Did the economic decision-maker attend a meeting?" was added to the opportunity record.

This structure made loss data both more honest and more comparable in the following quarters. If you'd like to see how an AI-powered CRM setup can be used across the entire pipeline, take a look at our case study on sales pipeline optimization with AI-powered CRM.

How did we automate the analysis with n8n and AI?

Manual analysis is valuable but not sustainable. Since it wasn't realistic to expect the team to comb through dozens of opportunities by hand every quarter, we automated the process with n8n.

The workflow we built

Trigger: When an opportunity moves to the "Lost" stage in HubSpot, the n8n workflow starts.
Data collection: The workflow pulls the emails, meeting notes, activity history and stage durations associated with the opportunity.
AI-powered summary and classification: A language model reads the collected text, produces a short summary of the opportunity and suggests a likely loss reason based on the taxonomy.
Mismatch check: If the reason selected by the rep differs from the one suggested by the model, the opportunity is added to the sales manager's review list.
Win/loss interview invitation: For opportunities above a certain value, the customer is sent an invitation to a short feedback interview a few days after the loss.
Weekly report: All data is consolidated in a dashboard and delivered to sales leaders as a weekly summary.

An important note here: The AI doesn't make decisions; it only flags. The model's suggestion doesn't replace human judgment. But it does surface inconsistencies that might otherwise be missed. In this case, opportunities where the rep had selected "price" but the decision-maker never appeared in the correspondence were flagged automatically this way.

Findings: what were the real reasons for the losses?

When we brought the three sources together (reclassified CRM data, the document review and the win/loss interviews), the picture was very different from the initial assumption. In the sample case, the breakdown looked roughly like this:

Process-related losses were the largest group. In particular, the decision-maker never being involved and discovery calls remaining superficial stood out.
Fit-related losses came second. Companies that didn't match the ideal customer profile, were too small or had a problem the product didn't solve were entering the pipeline from outbound lists.
No decision was higher than expected. Some customers hadn't chosen a competitor; they had chosen to do nothing. This showed that the urgency of change hadn't been established.
Genuinely price-related losses accounted for only a small share. Opportunities where price alone was decisive were far below the initial perception of "more than half."
Competition-related losses were concentrated around one particular competitor. In deals lost to this competitor, customers generally cited ease of integration as the reason.

The stage analysis was also instructive. A significant share of opportunities were lost not at the proposal stage but after the demo. The demo was delivered as a standard product tour rather than being built around the customer's own scenario.

What did we change after the findings?

Findings are only as valuable as the action they lead to. We assigned an owner and a concrete change to each main category.

For process-related losses

The discovery framework was overhauled. The sales team was given a discovery template that systematically uncovers the customer's business problem, the cost of that problem and the decision-making process.
A decision-maker rule was introduced. An opportunity could not move to the proposal stage until the economic decision-maker had attended at least one meeting.
The demo was redesigned. Every demo was built around three concrete use cases identified during discovery.

For sales methodology and pipeline discipline, Sales Corner's data-driven B2B sales approach provided the framework for this transformation.

For fit-related losses

The ICP was redefined. The ideal customer profile was narrowed by identifying the common characteristics of won opportunities (company size, industry, technology stack, team structure).
Outbound lists were re-filtered. Target lists were rebuilt and enriched according to the new ICP criteria using a B2B data platform such as Apollo.io.
Early disqualification criteria were added. The qualification questions SDRs ask in the first call were updated, preventing poor-fit opportunities from entering the pipeline in the first place.

For a broader look at how to narrow targeting on the outbound side, see SAAS Corner's outbound sales and lead generation service built specifically for B2B SaaS.

For value-related and no-decision losses

A business impact calculator was created. A simple ROI model was developed that customers could use with their own data at the proposal stage.
A "cost of doing nothing" message was added. A section quantifying the operational and financial cost of staying with the status quo was added to sales presentations.

We examined the impact of communicating value in numbers on sales teams in detail in our case study on AI-powered value selling and ROI growth.

For competition-related losses

A battlecard was created to address objections around integration.
Integration requests compiled from lost opportunities were regularly passed on to the product team.

What were the results?

The impact of the changes didn't show up overnight; it appeared gradually over several quarters. The main effects we observed were:

Pipeline volume dropped somewhat, but its quality improved significantly. Because non-ICP opportunities were filtered out early, the team worked with fewer but better-fit opportunities.
The win rate increased. The close rate was clearly higher, especially in opportunities where the decision-maker was involved in the process.
Discounting pressure eased. Because value was communicated more clearly, the "price" objection came up less often. Leadership also abandoned the across-the-board discount policy it had been planning.
The sales cycle got shorter. Because no time was wasted on poor-fit opportunities and the decision-maker was involved early, deals moved faster.

This last effect shows how closely intertwined loss analysis is with the sales cycle. We covered our approach to shortening the cycle separately in our case study on shortening the B2B sales cycle.

Perhaps the most important outcome was cultural. Loss analysis stopped being a "finding someone to blame" meeting and became a regular ritual through which sales, marketing and product teams learned together.

How do you get this process started at your own company?

You don't need to launch a major project to adapt this approach to your company. The 30-60-90 day plan below is a good starting point.

First 30 days: See where you stand

List the opportunities lost over the past two quarters.
Extract the stage at which each one was lost and its current loss label.
Read the notes and correspondence for 20–30 randomly selected opportunities and challenge the accuracy of each label.
Draft a two-tier loss reason taxonomy.

Days 31–60: Hear the customer's voice

Select a sample of lost opportunities and schedule independent win/loss interviews.
Keep interviewing until patterns start repeating.
Map the findings onto the taxonomy and draw up your first action list.

Days 61–90: Make the system permanent

Roll out required, two-tier loss reason fields in your CRM.
Set up an automation triggered at the moment of loss (data collection, summarization and interview invitations with a tool like n8n).
Launch a monthly win/loss review meeting and assign an owner to every action.

What are the most common mistakes?

Looking only at losses. If you don't analyze won opportunities too, you won't know what's working. You need both groups for comparison.
Blaming the rep. If the analysis turns into a performance review, the data quickly becomes corrupted. The focus should be on the process, not the person.
Creating too many categories. Thirty different loss reasons won't reveal any patterns. Keep the main categories limited.
Not sharing findings with product and marketing. Many loss reasons are issues the sales team can't solve on its own.
Treating it as a one-off project. Markets, competitors and customer needs change. Loss analysis should be an ongoing process.

Frequently asked questions

How much data do you need for loss reason analysis?

A few dozen lost opportunities are usually enough to see meaningful patterns. What matters more than data volume is that every opportunity is fully recorded with core fields such as stage, decision-maker involvement and competitor information. A small amount of high-quality data yields more valuable results than a large amount of superficial data.

Should sales reps conduct win/loss interviews?

Ideally, no. Customers find it hard to give critical feedback to the person who was trying to sell to them. An independent interviewer from sales operations, marketing or an external consultancy will collect much more honest answers.

Are opportunities tagged "price" really lost because of price?

Sometimes yes, but more often price is a symptom. When customers can't clearly see the value being offered, they find the price too high. Asking "If the price had been the same, would your decision have changed?" is a practical way to separate genuinely price-related losses from the rest.

How can AI be used in loss analysis?

AI can summarize the email, meeting note and activity data for lost opportunities and suggest a likely loss reason based on the taxonomy. Its most useful application is surfacing mismatches between the reason the rep selected and the reason the data points to. The final assessment should still be made by a human.

Conclusion

Finding out why you're losing B2B deals takes much more than looking at the labels in your CRM. In this case, behind what appeared to be a "price" problem were insufficient discovery, failure to reach the decision-maker, leads that didn't fit the ideal customer profile and value that was never made tangible. When a structured taxonomy, independent win/loss interviews, disciplined CRM fields and automations built with n8n came together, loss analysis stopped being guesswork and became a system that drives action. As a result, the company avoided an unnecessary discount policy.

If you're curious about the real reasons behind your own loss data, we can take a look together with our team, which combines expertise in sales, CRM and automation. Get in touch through the Corner Group contact page to talk through the process.

Lead Assignment Automation with n8n: Setting Up Round-Robin ›