---
title: The integration tax by team size: 10, 40 and 150 people
canonical: https://getcyril.com/blog/integration-tax-by-team-size/
published: 2026-10-12
author: Cyril team
language: en
---
# The integration tax by team size: 10, 40 and 150 people

[The integration tax](/blog/the-integration-tax/) priced one company: a 40-person business running sales, support, docs, projects and finance in five separate tools. The obvious next question is whether the shape holds for a ten-person firm, or for one with 150 people.

So here is the same model at three sizes. It is the arithmetic behind the [integration tax calculator](/tools/integration-tax/), with its default assumptions, and every figure below can be reproduced there by changing three inputs.

## How the three companies are built

The model has ten inputs. Six of them are unit prices and rates, and they stay at the calculator's defaults for all three companies:

- licences at a blended **$22 per seat per month**;
- AI add-ons at **$20 per seat per month**;
- **30 minutes per person per week** lost to reconciliation, at a blended **$34 an hour**;
- the person maintaining the seams on a **$70,000** fully loaded salary.

Three inputs follow headcount. The 40-person default has 128 paid seats and 52 seats with an AI add-on, which is 3.2 paid seats and 1.3 AI seats per person. The 10- and 150-person companies keep those ratios.

Two inputs are held at the 40-person default on purpose: **$400 a month** of connective software, and **0.35 of a full-time person** maintaining the seams. Neither of them grows in a straight line with headcount, and holding them still isolates the effect of size. The section after the table shows what happens when they move.

| Input | 10 people | 40 people | 150 people |
|---|---|---|---|
| People in the business | 10 | 40 | 150 |
| Total paid seats | 32 | 128 | 480 |
| Seats with an AI add-on | 13 | 52 | 195 |
| Everything else | Default | Default | Default |

The formulas are the calculator's own. Licences are seats times price. Connective software is a flat monthly amount. Administration is the FTE fraction times the salary, divided by twelve. Reconciliation is people times minutes per week, converted to hours, times 52/12 weeks a month, times the hourly rate. AI is seats times price.

## The integration tax at 10, 40 and 150 people

Monthly cost in US dollars, with each line's share of the total in brackets.

| Line | 10 people | 40 people | 150 people |
|---|---|---|---|
| 1. Licences | $704 (17%) | $2,816 (30%) | $10,560 (38%) |
| 2. Connective software | $400 (10%) | $400 (4%) | $400 (1%) |
| 3. Administration of the seams | $2,042 (49%) | $2,042 (22%) | $2,042 (7%) |
| 4. Reconciliation and error | $737 (18%) | $2,947 (32%) | $11,050 (40%) |
| 5. AI add-ons that see one silo | $260 (6%) | $1,040 (11%) | $3,900 (14%) |
| **Total per month** | **$4,142** | **$9,244** | **$27,952** |
| Total per year | $49,708 | $110,932 | $335,420 |
| Per person per year | $4,971 | $2,773 | $2,236 |
| Share that is not licences | 83% | 70% | 62% |

Source: the Cyril integration tax calculator at its default assumptions, with the headcount inputs above. An illustrative model, not survey data. Shares are rounded, so a column can sum to 99%.

The 40-person column is the original post's company. The post rounds each line for readability, so its table totals $9,230; the calculator does not round, which is where $9,244 comes from.

## Why the proportion falls, and where it stops

Three of the five lines grow with headcount: licences, reconciliation and AI add-ons. Two do not: connective software and the administration of the seams are, in this model, the same at any size.

At 10 people those two fixed lines are 59% of the total. Half the bill is a third of someone's time spent keeping five tools in step, and in a team that small that someone is usually a founder or a senior consultant. That is why a small team pays the most per person: $4,971 a year each, against $2,236 at 150 people.

As the team grows the fixed lines are spread thinner, so the share that is not licences falls: 83%, then 70%, then 62%. It does not fall towards zero. Per person per month, the three lines that grow cost:

- licences, $70.40 (3.2 seats at $22);
- reconciliation, $73.67 (30 minutes a week at $34 an hour);
- AI add-ons, $26 (1.3 seats at $20).

Reconciliation alone costs more per person than the licences. So however large the company, the share that is not licences in this model levels off near 59%, and the licences never reach half the bill. It is also why reconciliation, not licences, is the largest single line at 40 and at 150 people.

## The two places this table is most likely wrong

Holding connective software and administration still makes the comparison clean. It is probably not how real stacks behave, so here is what moving them does.

**At 10 people, 0.35 FTE may be generous.** Some small teams really do spend a third of a person on their tools; many spend less. Set administration to 0.15 FTE and the total falls to $2,976 a month. Licences are still only 24% of it.

**At 150 people, $400 and a third of a person are probably too low.** More people means more accounts to create and remove in five systems, and usually more connectors. Set administration to one full-time person and connective software to $1,200 a month, and the total rises to $32,543 a month. The share that is not licences goes back up to 68%.

In both directions, the conclusion of the original post holds: the subscriptions everyone negotiates are the minority of the cost.

## What the model leaves out

It is deliberately narrow, and it is worth being clear about what it does not count.

- **The questions nobody asks.** A cross-company question such as which accounts are at risk this quarter needs pipeline, tickets, delivery and invoices together. In five tools it is a project, so it gets asked rarely. The model gives that no value.
- **Switching costs.** Moving to anything else costs time and money. The model prices the stack you have, not the move.
- **Local prices.** The defaults are illustrative US dollar list-price assumptions. Substitute your own.

## Where Cyril's price sits against it

The calculator also shows Cyril's price for the same headcount, because Cyril is one of the things this tax can be compared against. Cyril is priced per seat with every app included: Starter is $29 a seat for 1 to 10 seats, Growth $59 for 11 to 50, and Business $99 for 51 to 200 ([see the plans](/pricing/)).

That gives $290 a month at 10 people, $2,360 at 40 and $14,850 at 150. The calculator's own rule is that Cyril's price replaces the licence line, the connective line goes to zero, and administration and reconciliation shrink rather than disappear, so it does not subtract them for you.

At 10 and 40 people, Cyril's price is below the licences and connective software it replaces ($1,104 and $3,216). At 150 people it is above them ($14,850 against $10,960), so at that size the case rests on the other three lines shrinking. Whether they shrink enough is a judgement about your own firm, and [the comparison of one platform against five tools](/compare/five-tools/) goes through it line by line, including where five tools remain the better answer.

## Reproduce any number in this post

1. Open the [integration tax calculator](/tools/integration-tax/) and reset it to its defaults.
2. Set people, total paid seats and seats with an AI add-on to one column of the inputs table.
3. Leave everything else alone. The result card shows the same lines, the same total and the same share.

The per-person row is the annual total divided by headcount. For the two adjusted scenarios, also change "Time spent maintaining it" and "Automation and connectors" as described above.

## Common questions

### Is this survey data?

No. It is an open model with every assumption stated, run at three sizes. No company was surveyed and no figure here was measured. If you cite it, cite it as a model.

### Why hold connective software and administration flat?

To isolate the effect of headcount. Both depend more on how many tools and connectors you run than on how many people you have, so scaling them in a straight line would overstate them at 150 people and understate them at 10. The section on where the table is most likely wrong shows what happens when they move.

### Why does a ten-person team pay more per person?

Because the fixed lines do not shrink with the team. Keeping five tools in step takes roughly the same effort whether ten or forty people use them, so at ten people that effort is spread across far fewer heads.

### Can I use this table in my own article?

Yes. Please describe it as an illustrative model rather than research, state that it uses the calculator's default assumptions, and link to this post or to the calculator so readers can change the inputs.

---

If you want to be among the first to use Cyril, [join the waitlist](/waitlist/).