How to spot sales bottlenecks by measuring each stage: wait times, attempts per lead, stalled leads, and root cause vs. symptom.
A sales bottleneck is the stage in the sales process where work piles up, waits, or is abandoned more than in the rest. Finding it does not require more intuition or more leads: you need to measure how long it takes for a lead to move from one stage to the next, how many attempts are made to move it forward, and what the team is doing in the meantime. This article explains the method in five steps, with examples you can replicate in your team.
Written by Pablo Pascual, co-founder of Vixiees. Updated in November 2026.
Key Takeaways - A bottleneck is seen in wait times and in the backlog of each stage, not in final revenue. - Measure each stage using three data points: time to the next action, attempts per lead, and stalled leads. - Separate the cause from the symptom before changing anything: "we are selling less" is almost never the real problem. - Fix only one stage at a time and measure again for two to three weeks before touching the next one.
What a sales bottleneck is
A sales bottleneck is the part of the sales process that limits the pace of the entire team. It can be in the first contact, in qualification, in the proposal, or in follow-up. Just like in a bottle, what goes in at the top does not come out faster than what the neck allows through. That is why a team can have many leads and few sales without the problem being lead generation.
The key is that the bottleneck is measured stage by stage. If you only look at the end result, you know something is failing, but not where.
This article proposes the four-readings-per-stage method: exit criteria, wait time, attempts per lead, and cause versus symptom. The following five steps apply these readings in order.
Why the symptom is misleading
When sales drop, the most common reaction is to look for more leads or change the offer. That reaction makes sense if the problem is demand, but not if the problem is execution. A team that takes two days to call a warm lead does not need more leads: it needs the lead to reach someone before it goes cold.
An illustrative example makes this clear. Imagine a team that receives 200 leads per month and closes 10. If 60% of those leads never receive a second call, the bottleneck is in follow-up, not in advertising. Without measuring by stage, that team will spend months testing new campaigns.
Step 1: define each stage with exit criteria
Before measuring, write down the stages as they actually happen in your team, not as they appear in the CRM template. For each one, define what needs to happen for a lead to leave it. Examples of exit criteria:
First contact: the person has been spoken with, not just had their number dialed.
Qualification: the need, timeline, and whether the person can decide or pay are known.
Proposal: a specific offer has been sent with a response deadline.
Closing: there is a signature, a payment, or a disqualification reason recorded.
If a stage does not have clear exit criteria, you will not be able to measure how much time is spent in it. That void is already a clue: it is usually a stage where the lead is left without an owner.
Step 2: measure the wait time in each stage
The most useful data point is the amount of time a lead spends in each stage before leaving. You can get it from your CRM, a spreadsheet, or manual tracking over a week. The important thing is to compare the median of each stage, not the average, because a few very old leads distort the average.
Pay attention to two things. First, which stage has the highest median. Second, what percentage of leads exceeds the timeframe that the team considers normal. For example, if the first contact should occur within an hour and 40% of leads take more than a day, you have a bottleneck candidate right there.
Queueing theory explains why wait time and the queue go together. John D. C. Little, professor at MIT Sloan, proved that the average number of items in a system is equal to the average arrival rate multiplied by the average time each one spends inside (L = λW). In his 2011 review in Operations Research, he puts it this way: "in most cases, service is the bottleneck that creates the queue."
An illustrative example applied to sales: if 50 leads enter the first contact stage per week and 100 are waiting, each lead spends an average of about two weeks in it. Little and Stephen Graves, also from MIT, write in a 2008 chapter on this law that average wait time and the average number of items waiting are important measurements for managers. The formula works with averages; to compare stages with each other, the median remains more reliable.
Oldroyd, McElheran, and Elkington summarize it in the blurb of their 2011 article in Harvard Business Review: "our research shows that most companies are not responding anywhere near fast enough." The study's exact numbers are behind a paywall and we do not quote them here, but the idea is clear: first contact speed is a stage that can be measured and improved.
Step 3: count attempts and actions per lead
Time is not enough. A lead might sit idle for a week because the sales rep had no defined cadence, or because they called twice and no one ever tried again. To distinguish between both cases, count three things per stage:
How many contact attempts are made before giving up on a lead.
What portion of leads has a scheduled next action.
How many leads are stalled, meaning without activity recorded in the last week.
This last figure is usually the most revealing. It is easy to discover dozens or hundreds of "open" leads that no one has touched for weeks. That is not a lead generation problem; it is a follow-up logjam.
Step 4: separate the cause from the symptom
Once you have the data, it is time to interpret it. Here, it helps to distinguish the symptom (what you see) from the cause (what produces it). Some common pairs:
Symptom: few proposals sent. Cause: qualification takes so long that leads arrive at the proposal stage already cold.
Symptom: many uncontacted leads. Cause: there is no clear distribution, and each sales rep waits for another to take it.
Symptom: sales drop during low season. Cause: the team stops following the same cadence when there is less demand. We explain this in low season in sales.
A simple test: if the stage you point to as the problem has normal times and the backlog is in the previous stage, the problem is not where you thought. Repeat the analysis one stage earlier.
Step 5: fix a single stage and measure again
Once you have the cause, change only what is necessary in that stage. For example, if the first contact takes too long, define an owner per lead and a distribution rule, and do not touch anything else for two or three weeks. That way, you will know if the change worked.
This order matters. If you change three stages at once, you will not know which change moved the needle. The one-change-at-a-time method also prevents the team from feeling like the process changes every week. If the change affects multiple reps, follow the order in implementing a sales process change: small pilot, comparison, and then the rest of the team.
This table summarizes what to look at in each stage. The alert timeframes are examples; adjust them with your team's data.
Stage | Exit Criteria | Metric to Measure | Sign of a Logjam (example) |
|---|---|---|---|
First contact | The person has been spoken with | Time to first call | Uncalled leads after more than a day |
Qualification | Need, timeline, and decision-maker known | Attempts per lead | Fewer than three attempts before disqualifying |
Proposal | Offer sent with response deadline | Days in stage | Proposals without a scheduled next action |
Closing | Signature, payment, or disqualification reason | Stalled leads | No activity in the last week |
How Vixiees helps you see it
Vixiees does not replace your CRM. It sits on top of daily work so that every lead has an owner, a next task with a specific time, and a record of what happened. With that data, the bottleneck stops being a gut feeling and becomes a metric for each stage. It does this with three features: it decides the next contact for each lead and records what happened, transcribes every call and automatically flags the ones that went poorly, and allows you to join an ongoing call silently to coach on the spot.
The Pontia case study, an online school for artificial intelligence and big data, shows the pattern. Jesús Garrido, CEO and founder of Pontia, explains that the problem was the wait time before the first contact. Previously, lead assignment was manual, and between 30 minutes and three hours could pass before someone picked up a lead. After automating assignment as soon as the lead comes in, the first contact no longer depended on who happened to be free.
According to Pontia's own statement in their customer case study (2025 period), daily leads per advisor became "almost double, with the same team." These are self-reported figures by the client, not independently verified. What does align with the method is the visibility aspect: dashboards allow you to see which stage is getting stuck and adjust the playbook using data.
A Vixiees client also shares this in their success story, From 1 to 8 accounts. In their words, previously they only saw how much they spent on advertising and how much revenue they generated, but not what was happening in the middle of the process. They describe it as being able to detect at which stage the roadblock preventing the business from scaling is located. It is their experience, not an audited measurement.
For the method to work, the team needs three things: stages with exit criteria, a visible owner for each lead, and an activity log that updates without asking for permission. If one is missing, the figures will be incomplete and the bottleneck will remain hidden. You can read more about how a lead is distributed among reps in our article on duplicate leads, and on how to sustain the process when a team member is away in leads when a sales rep goes on vacation.
Common mistakes when looking for bottlenecks
Most mistakes come from measuring the end result instead of each stage, or changing too much at once. These are the five most common:
Measuring only revenue. It tells you that something is failing, but not at which stage.
Using the average instead of the median. A few old leads hide the true pattern.
Counting leads instead of actions. A lead with five recorded attempts is not the same as one with zero.
Changing several things at once. It prevents you from knowing what worked.
Blaming the person before reviewing the process. If the stage lacks exit criteria, anyone will take too long to move it forward.
If the process is already well-defined and the numbers still do not add up, it is worth reviewing the offer and pricing side as well, because sometimes the bottleneck lies outside the team.
When a team tells me they're not selling enough, the first thing I look at isn't the final result, but at what stage the leads are getting stuck. Almost always, the bottleneck occurs before the negotiation: in how long it takes to reply, in who is assigned the lead, or in how many times contact is attempted. My advice is to choose a single stage, measure how long each case takes for two weeks, and change just one thing. If you change three things at once, you won't know which one worked.
Frequently Asked Questions
How do I know if I have a bottleneck in sales?
Compare the time a lead spends in each stage and look for the one with the highest median or the most stalled leads. If one stage concentrates the wait time, that’s your candidate. Confirm it by tracking metrics for two or three weeks before changing anything.
What data do I need to find the bottleneck?
You need the entry date for each stage, the date of the last action, and the outcome. With that, you can calculate times, attempts, and stalled leads. You don't need a complex system: a spreadsheet with one row per lead is enough to get started.
Is the bottleneck always in the sales team?
No. It can be in lead generation, if leads that don't fit the offering are coming in, or in the product, if the proposal takes too long to decide on. That's why it's best to start with the measured stages and work backward if the initial contact goes well.
How long should it take for a lead to move through each stage?
There is no universal number. It depends on the sales cycle and the product. Define your team's typical timeframe using their data from the past few weeks, and flag any lead that exceeds it as an alert.
Do I need to switch CRMs to fix the bottleneck?
Often, no. Usually, what's missing is an exit criterion, an owner, or an assignment rule. A well-configured CRM can give you the numbers, even though changing the process is what truly moves the needle.
If you want to review your own stage-by-stage process, you can speak with the Vixiees team and we'll measure it together before changing anything.
Sources
John D. C. Little (MIT Sloan). "Little's Law as Viewed on Its 50th Anniversary". Operations Research, vol. 59, no. 3, 2011, pp. 536-549. DOI 10.1287/opre.1110.0940. RePEc listing.
John D. C. Little and Stephen C. Graves (MIT). "Little's Law", chapter 5 of Building Intuition: Insights From Basic Operations Management Models and Principles. Springer, 2008. Springer.
James B. Oldroyd, Kristina McElheran, and David Elkington. "The Short Life of Online Sales Leads". Harvard Business Review, March 2011. HBR. Qualitative use: the public version does not display the figures.
Pontia. Customer case study on Vixiees, 2025 data reported by the customer. vixiees.com/customers/pontia-tech.
"From 1 to 8 accounts" case study, Vixiees customer on video. YouTube. Customer experience, no figures used in the article.
