Your reps already move deals through stages. What they usually can't see, at a glance, is which stage is clogging and which deals have gone cold sitting in it. A kanban board fixes exactly that problem: a work-tracking system built around one idea, that anyone should be able to see where every deal sits and where it has stalled, without pulling a report or interrogating a rep. Generic guides teach this to software teams. This one translates every mechanic into pipeline decisions a sales manager actually makes.
What a Kanban Board Actually Is (and What It Is Not)
Strip a kanban board down and you get three parts. Columns are the stages work moves through, left to right. Cards are the individual units of work, one card per item. Limits are caps on how many cards a column may hold at once. That third part is what separates a real board from a whiteboard covered in sticky notes.
A to-do list tells you what exists. A kanban board tells you what is moving, what is stuck, and where the pressure is building. That difference matters because a sales pipeline is a flow system: deals enter, advance when they are ready, and exit as won or lost. Visualizing that flow is the board's job, not storing the deal record. Act on it today, not at quarter close.
The approach traces back to Toyota, where engineer Taiichi Ohno used physical cards (kanban means 'visual signal' in Japanese) to signal when a station was ready to pull more work. The core idea borrowed for sales: pull, not push.
How a Kanban Board Maps to Your Sales Pipeline
The most common mistake teams make when they build their first board is labeling columns by activity: Calls, Emails, Demos, Follow-ups. It feels organized. It is wrong. Columns are stages a deal passes through once, in order, not tasks a rep repeats. A deal does not move from 'Emails' to 'Calls'; it moves from Qualified to Discovery Scheduled. Activity-type columns turn the board into a task list and destroy the one thing it was built to show you: forward progress.
Cards are deals, and what you put on them matters
Each card is one deal, and what you print on it decides whether the board is glanceable or useless. Three things earn their place: the account name and deal value, so you know the stakes; a days-in-stage counter, so you can see aging; and the next committed action with its date, so you know whether the deal is actually alive. A card crammed with every CRM field is as unreadable as the spreadsheet you were trying to escape. In KudosCRM's kanban view, the deal card is configurable, so you can strip it to those few fields instead of recreating a grid.
Pull, not push: why reps move deals, not managers
In a pull system, a rep advances a deal into the next stage when they have the capacity and the deal has met that stage's entry criteria. Push works the opposite way: a manager drags deals forward to hit activity targets, or reps advance deals early to look busy. Push inflates every downstream stage with deals that are not ready, which is how a Proposal Sent column fills with proposals nobody asked for.
Pull keeps the board honest. A deal advances because it earned the move, not because someone needed a number to look better this week.
Designing Your Sales Board Column by Column
Six columns cover most B2B sales processes without turning the board into CRM bureaucracy. Below is a layout that works for mid-market teams, with a starting WIP cap and the two card fields that carry the most weight in each stage.
Column | Starting WIP (per rep) | Show on the card |
|---|---|---|
New Lead | no cap | Source, lead score |
Qualified | 6 | Account, why it fits |
Discovery Scheduled | 6 | Meeting date, days in stage |
Proposal Sent | 8 | Deal value, days since sent |
Negotiation | 4 | Deal value, next committed step |
Closed Won | no cap | Deal value, close date |
New Lead and Closed Won are holding zones, not active work stages, so they stay uncapped. The four stages between them are where deals consume rep attention, and those are the ones worth limiting.
If you catch yourself wanting an eighth or ninth column, ask whether it reflects how deals actually move or just a reporting field your CRM asked you to fill in. Seven columns is a soft ceiling. Past that, the board mirrors bureaucracy instead of flow.
WIP Limits: How Many Deals Per Stage Before Quality Drops
Picture a rep holding eleven live proposals. Each buyer expects a tailored quote, a prompt follow-up, and answers inside a day. At eleven, the rep is pasting last week's proposal and calling it personalization. Quote quality slips, and so does the close rate on exactly the deals that were closest to money.
A work-in-progress limit caps how many deals one rep can hold in a stage at once, set to the number a rep can genuinely give attention to. That number falls as the effort per deal rises. A 2026 Bitrix24 analysis puts sensible per-rep caps at 6 to 8 for discovery, 4 to 6 for proposals, and 3 to 5 for negotiation, because a live negotiation demands far more of a rep than a discovery call does. Start there and tune to your team's real throughput.
The capacity ceiling: WIP cap times rep count
Multiply the per-rep cap by your headcount and you get the team's ceiling for that stage, a hard limit visible on the board rather than buried in a report.
Cap Proposal Sent at 8 deals per rep. Run 5 reps. The board can hold at most 40 active proposals at once (8 x 5 = 40). When Proposal Sent sits at 38, you know the stage is about to jam before a single deal slips, and you can act: push the oldest proposals to a decision or clear them out. Most managers find a stalled stage only after the quarter misses. The ceiling math lets you watch the jam form while there is still time to fix it.
Cycle time falls out of the board for free
Cycle time, the average number of days a deal spends in a stage, is usually a separate report nobody runs. On a board with a days-in-stage counter on every card, it is just the board's memory. Watch how long cards sit in Negotiation before they move, and you have the metric without building anything.
For reference, mid-market B2B deals spend roughly 7 to 14 days in discovery and 14 to 30 in negotiation, per Zeliq's 2026 cycle-time benchmarks, and the median mid-market cycle now runs about 92 days, up from 68 in 2019. When your board's live numbers run well past those ranges, that stage is where your pipeline is losing time.
Reading Board Aging to Catch Stale Deals Early
A deal sitting in Discovery Scheduled for ten days with no logged activity is not slow. It is dying, and the board should flag it in a color you cannot ignore. Aging is the single most actionable signal a kanban board produces, but only when thresholds are set per stage rather than a blanket 14 days across the whole board.
What counts as stuck changes by stage
Discovery Scheduled with no logged activity after 10 days means the meeting never happened or never got booked, and the deal is cooling fast. Proposal Sent can reasonably age to 14 days, because buyers take time to review, but past that with no response, it needs a nudge or a disqualification. Negotiation is the tightest: 7 days without a documented next step means the deal has gone quiet at its most fragile point, and quiet in negotiation usually means you are being shopped or ghosted.
Same board, three different clocks, because a healthy negotiation moves faster than a healthy discovery.
Three questions for any deal past its threshold
When a card crosses its threshold, a short set of questions sorts it fast.
First, is there a real, dated next step, or just a hope? A deal with no scheduled action is not in the pipeline, it is in limbo. Second, has the buyer done anything in the last week: opened the proposal, replied, booked time? Silence on their side is the tell. Third, would you bet your own money this closes this quarter? If the honest answer is no, move it back a stage or out of the pipeline entirely, and stop letting it inflate your forecast.
Kanban Board vs. CRM Pipeline View
These get treated as the same thing, and they are not. A CRM pipeline view is a record system: a grid or list where you enter deal data, run reports, and pull the numbers finance wants. Its job is storage and reporting. A kanban board is a flow visualization that shows the live state of every deal in motion, surfaces which stage is jammed, and makes aging visible on the card. Its job is to help you act today.
The pipeline view answers one question: what do we have, and what is it worth? The board answers a different one: what is moving, what is stuck, and where do I step in?
You need both. Reach for the grid when you are reporting up or reconciling numbers. Reach for the board when you are running the team, triaging deals, or deciding where a rep should spend the next hour. KudosCRM's kanban pipeline view maps your existing deal stages into columns, puts an aging counter on every card, and shows the stage that is about to jam before it costs you deals. It is the same pipeline data you already keep, shown as flow instead of a grid. For a deeper look at structuring the stages themselves, the guide to sales pipeline stages pairs naturally with the board design above.