An MSP owner opens the forecast on the first of the month. It says four hundred thousand dollars.
Inside that number: two managed agreement renewals, a server refresh for a client who has already said yes and is waiting on stock, a security uplift project that has been "nearly signed" since June, and a new logo that came through a referral three weeks ago.
Five things. They close on timescales ranging from a fortnight to nine months. One of them is mostly the cost of hardware passing through the business. Two of them, if they land, keep paying every month for the next three years. One of them is a renewal of revenue the business already has, which means if it closes nothing changes and if it does not, something quite bad happens.
The forecast adds them up and says four hundred thousand dollars.
That number is not wrong because somebody was optimistic. It is wrong because it is a sum of things that do not belong in the same sum.
The three lines, and how differently they behave
Almost every managed services business is three businesses wearing one logo.
| Line | What it is | How it closes | What it does after it closes |
|---|---|---|---|
| Managed agreements | The contracted monthly services. The reason the business is valued the way it is | Slowly, often against an incumbent, usually with a procurement step | Keeps paying every month for the term. The value is a rate, not an amount |
| Projects and professional services | Migrations, rollouts, security uplifts, office moves. One-off work with a scope | Faster, often from inside an existing account, gated on budget rather than trust | Pays once, then is gone. Capacity is the constraint, not demand |
| Product | Hardware and licence resale | Fastest of the three, and frequently decided before you are asked | Pays once and mostly passes straight through. The profit is in the labour attached to it |
Industry benchmarking puts these in very different margin territory. Service Leadership figures reported for 2026 have professional services often targeted around 45 to 60 per cent gross margin, break and fix technical services typically at 30 to 40 per cent when priced properly and easily under 20 per cent when labour efficiency slips. ConnectWise and BrightGauge figures put mature managed services consistently at 50 to 60 per cent. Hardware sits below all of it, because you are reselling somebody else's product.
Which produces the trap. The line with the biggest invoice value has the smallest margin, and it is the one that most distorts a combined forecast. A ninety thousand dollar hardware order and a ninety thousand dollar project are the same number in a pipeline and nothing like the same event in the business.
What one pipeline does to a weighted forecast
A weighted forecast multiplies deal value by the probability attached to its stage. That arithmetic is only meaningful when the deals in the pipeline behave alike, because the probability of a stage is really a statement about what usually happens next to deals of that kind.
Put three kinds in together and the probabilities stop describing anything. A hardware order at "quoted" is close to certain, because the client asked for the quote and the decision is already made. A managed agreement at "quoted" is a long way from certain, because a quote is where the incumbent comparison starts. The same stage, the same percentage, two completely different realities.
The result is a forecast that is not cautiously wrong in one direction. It is wrong in a way that changes shape month to month depending on the mix, which is worse, because you cannot learn to correct for it.
The second effect is subtler. Because hardware and projects close faster, they dominate the near end of a combined pipeline. Agreements, which are the only line that compounds, sit further out and look sparse by comparison. A business optimising against that view will quietly prioritise the revenue that does not recur, which is the exact opposite of what its valuation depends on.
Forecast recurring revenue as a rate, not an amount
The most common mistake is putting a managed agreement in the pipeline at the total contract value and letting it sit beside a project at its one-off value.
A three-year agreement at eight thousand a month is not a two hundred and eighty-eight thousand dollar deal in the sense that a two hundred and eighty-eight thousand dollar project is. It is a change to the monthly rate the business earns, and the useful question about it is not "how much" but "what does this do to monthly recurring revenue, and for how long".
So forecast the agreement line as movement in recurring revenue: what is being added, what is at risk, what is churning. Forecast projects as value landing in a period. Show both, side by side, and resist the urge to produce a single total. The total is the number people ask for and the one that tells them least.
The revenue nobody sold
Here is the line item that does not appear in any of the three pipelines and probably should.
A client grows from thirty seats to fifty. Nobody ran a sales process. Somebody in the service desk noticed, or did not. The agreement gets updated, or it does not, and quite often it gets updated three months late, which is three months of unbilled seats and a slightly awkward conversation.
Growth inside the existing base is the cheapest revenue an MSP will ever earn, and it is invisible precisely because it never entered a pipeline. There is no owner, no value, no date and no forecast line, so there is nothing to miss. The business only finds out it did not happen when it does not happen.
The fix is to treat it like the other three: seat growth, a security uplift at renewal, a hardware refresh cycle that is due. Deals against existing accounts, with an owner and a date, worked like anything else. Most MSPs find their easiest quarter sitting in their own client list, and the reason they have not had it yet is administrative rather than commercial.
The renewal is a calendar, not a stage
One more distinction worth holding.
A renewal does not behave like a deal until something goes wrong with it. It has a date set by a contract rather than by a buying process, and in the normal case it rolls. Putting it in a pipeline at stage one, twelve months out, produces a forecast full of revenue you already have.
What renewals need is a calendar that starts a commercial conversation well before the date, so that the first sign of trouble is not a procurement email asking for a market comparison. The moment a renewal is contested it stops being a calendar item and becomes a deal, with a competitor, a decision maker and a real chance of loss. That is when it belongs in the pipeline.
Treating every renewal as a deal makes the forecast meaningless. Treating no renewal as a deal makes them arrive as a surprise. The line between the two is whether anyone else is being asked to quote.
What this asks of the software
Three pipelines with their own stages and their own probabilities, a view of recurring revenue that is separate from one-off value, deals attached to existing accounts and not only to new logos, and a renewal calendar that is not just a stage called "renewal".
That is a fairly ordinary set of requirements and a surprising number of tools cannot do it, because they assume one pipeline with one set of stages and a single weighted total. Empiraa Signal was built with this shape in mind: agreements, projects and new logos each get their own pipeline and probabilities, the forecast adds them up and puts them beside the budget for the year rather than collapsing them into one figure, and deals hang off accounts so growth inside the base is worked rather than noticed.
The pricing point matters here for a specific reason. Signal is A$99 a month including GST for the whole workspace with unlimited users on every plan, and in an MSP that is not a cost argument so much as a visibility one. The person who first notices a client has hired fifteen people is usually on the service desk, not in sales. Per-seat pricing gives you a reason not to give that person a login, which is how seat growth stays invisible.
The first thing to do
Do not reorganise anything yet. Take the current forecast and split it into the three lines on paper.
Most MSP owners doing this for the first time find the same two things. The near-term number is more hardware-weighted than they thought, so the quarter looks stronger than the gross profit will be. And the agreement line, the only one that compounds, has fewer things in it than the business needs to grow at the rate the owner is assuming.
Both of those are useful to know in September rather than in December. Neither of them is visible in a single total.
Related reading: sales pipeline management for growing teams, sales pipeline forecasting for growing teams, and the warmest pipeline you have is already in your CRM.
