CRM for a sales team: pipeline, plan vs. actual, conversion and stalled deals

A CRM for a sales team is a system in which every deal moves through pipeline stages with verifiable transition criteria. From the same data, the head of sales sees plan attainment, conversion and time in each stage, along with a list of stuck deals. Four things set it apart from a notebook with reminders. Stages describe events on the client's side: requirements received, proposal sent, meeting scheduled with the decision-maker. Actual sales come from payments in the accounting system. The forecast is calculated from the company's own conversion history. The system finds stalled deals itself, by rules, without polling reps every day. Popular CRMs have similar feature sets, so the system is worth choosing by how your pipeline fits into it.

There is no CRM ranking with pricing plans here. Below is what the system needs to do for a sales team and how to check it at a demo. How to set stages and transition criteria for your cycle: a short B2B sale, a long project deal, a tender. How to calculate conversion and time in stages, how to build plan vs. actual and a revenue forecast from the pipeline, and how to spot stalled deals without micromanagement. Separately: repeat sales to the existing customer base and the reports the head of sales receives without assembling them by hand. At the end: what to prepare before you start and why an off-the-shelf product covers only part of the job. Choosing a system in general is covered in the article on how to choose a CRM, and the deal's path from inquiry to payment, along with the link to 1C, the accounting and ERP platform most Russian companies run on, in the article “Sales automation”. This article is about how the head of sales runs the team through the pipeline.

What does a CRM need to do for a sales team?

A CRM for a sales team has to store each deal's path through the stages with transition dates, check the conditions for moving between stages, and calculate plan vs. actual, conversion and the forecast from that data. Any system can keep contacts and assign tasks. The difference shows in the head of sales' reports: whether they are built from the system's data or a rep assembles them in a spreadsheet before the team meeting.

The requirements are easiest to check at a demo using one of your own deals:

The last rows of the table usually decide whether the head of sales will live in the CRM or in separate files next to it. Hardly anyone drops the other tools alongside the CRM. J'son & Partners Consulting surveyed a thousand companies from all of Russia's federal districts between November 2025 and January 2026. According to its data, 98.7% of companies use CRM alongside other tools, most often office software. One in five companies, 20%, adopted a CRM after they started losing deal data. The results were published by the journal of Bitrix24, a Russian CRM and collaboration suite, on August 19, 2026, which makes it a CRM vendor's publication, and the source is worth keeping in mind.

How do you set pipeline stages and transition criteria for your cycle?

A pipeline stage is an event on the client's side that can be verified from the deal record, and a transition criterion is what has to appear in the record for the deal to move on. Stages called “in progress” and “thinking it over” measure nothing. The system can check stages like “requirements received” and “proposal sent, next contact date set” on its own.

The sales cycle works differently from one business to another, and the pipeline follows it:

A few rules are worth following when you set up stages:

1. Five to eight stages per pipeline as a guideline. In practice, that number is usually enough. Fewer, and different deal states get lumped into one stage. More, and reps start skipping stages, so the time spent in them stops meaning anything.
2. The stage name describes a client action or a fact. “Client received the proposal” can be checked; “Working on it” cannot.
3. Required fields at each transition. A deal cannot move to approval without an amount, an expected close date and a decision-maker contact. Keep required fields few, or reps will start filling them with anything.
4. Loss with a reason from a short list. Price, timing, chose a competitor, project postponed, no budget, not our client. A free-text “comment” field does not work for reasons: you cannot build a report on it.
5. The decision to bid on a tender as a separate stage. Record who decided to bid and why. Log a loss with the winning price. Six months later, you can see which tenders the company wins and which ones eat up the team's time with no real chance.

If each rep defines the stages their own way, you will not be able to compare their conversion. So the pipeline is described by the head of sales together with two or three experienced reps, not by each person for themselves. Why a pipeline that adds work for reps does not take hold is covered in the article “Why CRM doesn't take hold”.

How do you calculate conversion and time in each stage?

Stage conversion is the share of deals that moved on to the next stage out of the deals that entered this stage. Time in stage is how many days a deal spent between entering and leaving it. Both metrics are calculated from the transition history, so the CRM has to store the date of every transition, not just the deal's current stage.

A hypothetical example. Over a quarter, two hundred inquiries entered the pipeline:

There is no median in the total row: stage medians do not add up to the median of the whole cycle, which is calculated separately from closed deals.

A table like this shows where to look: the biggest drop and the longest wait both fall on the move from proposal to approval. Next, you open the loss reasons at this stage and the deals stuck in it. If half the losses are “price,” the conversation is about pricing policy. If deals are simply sitting with no response from the client, the conversation is about how reps follow up after sending a proposal.

Conversion can be calculated in two ways, and for a long cycle the difference matters:

• Per period. How many deals moved from each stage to the next during the month. This suits short sales, where a deal goes through the pipeline in a week or two.
• By cohort. Of the deals created in March, how many reached the proposal, approval and payment. For project sales with a cycle of several months, a per-period calculation mixes deals from different months, and conversion jumps around for no reason.

Time in stage is better viewed as a median: a single deal that sat in a stage for six months skews the average considerably. The median for each stage also becomes the norm for the stalled-deal rules discussed below.

The figures in such a report are only as accurate as the deals are moved carefully. On February 3, 2026, Salesforce published its State of Sales report: 4,050 sales professionals from 22 countries were surveyed in August and September 2025; Russia was not in the sample. According to the report, 74% of sellers are focused on cleaning up data. Among top performers, meaning those whose revenue grew noticeably over the year, 79% consider clean data a priority, compared with 54% of underperformers. If a rep moves deals between stages once a month right before the report, time in stage and cohort conversion will describe the rep's habits, not the sales.

How do you build plan vs. actual and a revenue forecast from the pipeline?

Plan vs. actual comes together when the plan is entered in the CRM by rep and by period, and actuals arrive from the accounting system as payments. The forecast is built from the deals currently in the pipeline, taking into account their stage, amount, expected date and the company's own conversion history. After that, the spreadsheet for the team meeting is no longer needed: the figures update along with the data.

First, agree on what counts as actual. The most reliable measure is money received; in some companies it is a shipment or a signed acceptance certificate. The “deal won” mark in the CRM should not count as actual: a rep sets it when the client says “we'll take it,” and the money arrives a month later or not at all. Payments are recorded in accounting and reach the CRM through data exchange with 1C. How this works is covered in the article “Sales automation”.

The plan is broken down the same way the actuals will later be read:

• By rep and by team. Everyone sees their own plan and actuals as of today; the head of sales sees them by team.
• By month within the quarter. A shortfall shows up mid-month, while there is still time to catch up.
• By new and existing clients. Hitting the plan thanks to one large regular client and hitting it through new deals are different situations.
• By business line or product, if their margins differ.

The forecast is built from two parts, because each one on its own gets it wrong.

1. Weighted pipeline. Each deal's amount is multiplied by the probability of its stage, and the results are summed by expected close date. The probability comes from your own conversion history: in the example above, 60% of deals go from approval to payment, so a deal at that stage counts for 60% of its amount. The default probabilities in the CRM settings have nothing to do with your sales.
2. The rep's assessment. The rep places each deal for the current month into one of three categories: “certain,” “likely,” “in doubt.” The head of sales compares this assessment with the weighted pipeline and reviews the deals where the two differ sharply.

The same history tells you how many deals the pipeline needs to hit the plan. Take the same hypothetical table: 43% of deals go from proposal to approval, and 60% from approval to payment, so roughly one in four goes from proposal to payment. Say the quarterly plan is ten million rubles and deal amounts are similar to each other. Then you need about forty million in deals at the proposal stage. If there are twenty, the plan is at risk as early as the first month of the quarter.

Sales leaders put little trust in sales forecasts. On July 19, 2022, Gong published its Reality of Forecasting report: 928 sales and revenue professionals were surveyed; the release does not give the survey dates or countries. 24% of sales leaders trust the forecast their reps give them. 52% say their company regularly misses its forecast by more than 10%. Gong sells sales analytics tools, and the source is worth keeping in mind. The conclusion itself can be checked against your own data: compare the forecast with actuals every month and see at which stage the probabilities are inflated.

How do you spot stalled deals without micromanagement?

Stalled deals are found by rules: the CRM itself flags a deal with no next action, a deal that has sat in a stage longer than the norm, and a deal whose close date has been pushed back repeatedly. The rep gets the signal first, and only if they do nothing does the deal go to the head of sales. The head of sales looks at a list of exceptions, not at the entire pipeline.

The norms for these rules come from your own data: the median time per stage, the usual rhythm of contact with clients, the time it takes to prepare tender bids. Other companies' norms will not help here, because selling consumables and selling equipment move at different speeds.

The “silence in communications” rule only works if calls and emails reach the deal from telephony and email on their own. When they have to be logged by hand, the system catches not stalled deals but reps who forgot to make an entry, and trust in the rules quickly fades.

This changes how the head of sales talks to the team. The team meeting covers five to seven deals from the exception list: what is blocking them, who is needed on the client's side, close or move forward. Reps whose deals are moving do not end up in the review and do not spend time recounting their pipeline. Calls and emails as a performance metric stay in the background: they show that a rep is working, but not whether deals are moving.

How do you handle repeat sales and upselling to your customer base?

Repeat sales are handled in a separate pipeline, and the system finds the triggers for them in the purchase history from the accounting system. A client has missed their usual ordering window, a contract or limit is running out, a client buys one product group but not the related one. Each such trigger becomes a task for the rep with a clear reason to call.

Mixing repeat sales with new ones is inconvenient: they have different stages, different conversion and a different cycle length. If everything is in one pipeline, the high conversion of regular clients masks failures in acquiring new ones, and vice versa.

Triggers for a repeat sale that the CRM finds on its own:

• A longer pause than usual. A client orders once a month, but the last shipment was two months ago. The rhythm is calculated from shipment history in 1C, without manual tracking.
• A contract, limit or subscription ending. A set period before it ends, the rep gets a task to discuss renewal.
• An incomplete basket. A client buys the main item but not the consumables and components usually bought with it.
• A drop in volume. The client's order total for the quarter is noticeably lower than for the same quarter a year ago.
• A change of contact at the client. The buyer you worked with has left. It is worth introducing yourself to the new person before a competitor does.

For these rules to work, the customer base has to be clean: one client, one record, with all shipments linked to it. How to put the base in order and keep clients when a rep leaves is covered in the article “The customer base in a CRM”. In manufacturing, a repeat order is more complicated, because it is almost always a variation on a previous specification; this is covered in the article on CRM for manufacturing.

Which reports does the head of sales receive automatically?

The head of sales automatically receives the reports whose data the system collects on its own: plan vs. actual, the pipeline with conversion and time in stages, the forecast, the list of stalled deals, loss reasons and repeat sales. If even one figure in a report is entered by hand before the team meeting, the report is no longer automatic and starts to lag behind reality.

The “lead sources through to payment” report often changes how people see their channels. Sometimes the channel with the most inquiries produces few payments, and without the link to payments this stays invisible. The “forecast accuracy” report is needed so that in two or three months people start trusting the forecast.

Reports are easier to receive on a schedule by email or messenger than to look for in the interface; this is how the “Scheduled auto-reports” and “Summary dashboards” solutions from the catalog work. Which sales figures belong on the owner's screen and which to leave to the head of sales is covered in the article “The owner's dashboard”. How sales tie into the rest of management reporting is covered in the article on management reporting.

Where do you start?

Start by describing the pipeline on paper: the stages for each type of sale, the transition criteria, the time norms and what counts as actual. This description shows what CRM configuration will cover and what will require a link to accounting and additional development. The company does most of the preparation itself.

1. Documented pipelines. The stages for each type of sale, the transition criterion for each stage and the required fields. Write down how you actually sell, including the workarounds through messengers and personal spreadsheets.
2. A list of loss reasons and rules for closing deals. A short list of reasons and a deadline after which a deal with no movement is closed as lost rather than kept in the pipeline “just in case.”
3. The plan, broken down. By rep, month, new and existing clients, business line. The head of sales decides on the breakdown together with the owner.
4. Access to the accounting system. A separate 1C account with the necessary permissions and a copy of the database for testing the data exchange. Without payment data from accounting, plan vs. actual will end up in a spreadsheet again. How systems are connected is covered in the article on system integration.
5. Connected channels. Telephony, email, messengers and website forms, so that inquiries and communications reach the deal without manual entry. Without this, the stalled-deal rules work blind.
6. A permissions matrix. Who sees other people's deals, who changes the plan, who can close a deal as lost, who sees margin. Roles are easier to define before configuration than to fix after complaints.
7. An owner after launch, and ongoing support. An employee who looks after the stages, norms and reference lists, and a support contract for the data exchange with accounting. The sales team, products and process change, and without an owner the pipeline will drift away from how you actually sell within six months.

The first three items require no budget and on their own give the head of sales more clarity than switching systems. The other four determine the project's cost: how many channels and systems need to be connected, how many pipelines and rules need to be configured, and how many exceptions there are in the process. That is why the cost of a CRM for a sales team built around your process is quoted after an assessment.

Off-the-shelf or a system built around your process?

An off-the-shelf CRM covers the common part well: taking inquiries, client records, tasks, the pipeline and standard reports on it. The rules around the pipeline are specific to each company: procedures, reference lists, access rights and exceptions. These have to be configured or built on top.

At the start, sales teams look alike: an inquiry has to be received, an owner assigned, a proposal sent, and a follow-up call not forgotten. It makes sense to take this part ready-made, and for a small team with a single pipeline, a configured off-the-shelf product is often enough for a long time.

Companies diverge where the pipeline meets money and people. For one, actual means payment; for another, a signed acceptance certificate; for a third, shipment in installments. One runs tenders with bid security and deadlines from the notice; another sells under framework contracts with limits. At one company, the team lead sees margin; at another, only revenue. In an off-the-shelf product, these rules get forced into someone else's data model: through customizations that have to be checked after every update, or through manual reconciliations that fall on the head of sales. We covered the choice between a ready-made solution and a system built around your process separately in custom system or off-the-shelf.

In the project for a property developer, qualified inquiries from a first-line bot go into the CRM, where reps handle the deal. The CRM shows contact history, statuses, properties and the sales plan, documents are generated automatically, and the owners see the sales picture on dashboards. In the project for a wholesale and manufacturing company, sales are linked to accounting: the product catalog, stock levels, shipments and payments sync with 1C in both directions.

INCUBE AI takes on both parts of the job. For standard areas, there is a solutions catalog, and each solution is adapted to the client's way of working. A pipeline with your stages, plan vs. actual based on payments from 1C, rules for stalled deals and reports for the head of sales we build as custom development. We work under contract, data is stored in Russia, and after launch the system stays under support. If your plan vs. actual is still put together in a spreadsheet and the forecast doesn't match payments, tell us about your task: we will go through your pipeline, see where deals are being lost, and estimate the scope of work based on that review.

Sources and links

• J'son & Partners Consulting: how companies choose a CRM in 2026. Bitrix24 journal, August 19, 2026, by Marina Naumenko — a survey of a thousand companies from all of Russia's federal districts, November 2025 to January 2026. 98.7% of companies use CRM alongside other tools, most often office software. 20% adopted a CRM because they had started losing deal data. The results were published by a CRM vendor's publication.
• Salesforce: State of Sales Report, February 3, 2026 — a survey of 4,050 sales professionals in 22 countries, August and September 2025. 74% are focused on cleaning up data. 79% of top-performing sellers and 54% of underperformers consider clean data a priority.
• Gong: Reality of Forecasting Report, July 19, 2022 — a survey of 928 sales and revenue professionals. 24% of sales leaders trust their reps' forecasts. 52% say their company regularly misses its forecast by more than 10%. Gong sells sales analytics tools.