Section 1
You are selling a non-event, so you are designing a comparison
Picture a 40-seat accounting firm on a managed services contract. In a good month the client receives an invoice and experiences nothing at all. No ransomware, no three-day outage, no quarter-end close blown up by a failed server. That absence is the product. It is also the problem, because absence has no texture. The client cannot hold it, count it, or compare it against the absence a competitor would have produced for less. Economists call this information asymmetry: one side of a transaction knows something the other side cannot verify. Your buyer cannot assess whether your monitoring is tuned or noisy. They can assess a number. Hand them a document where the number is the only field that varies in a way they understand, and you have instructed them to buy on price. This is where signaling matters. A signal is something you do that is expensive for a weak provider to copy. Certification logos in a footer are cheap, every competitor has them, and they separate nobody from anybody. A response commitment with a service credit attached is costly, because a firm with a disorganised dispatch process cannot afford to make it. That gap between cheap and costly is the whole value of a signal. Packaging is how you put costly signals in front of a buyer in comparable form. Which makes it mechanism design rather than copywriting. Mechanism design means arranging the rules of a choice so that people acting in their own interest land on the outcome you want. Applied here: if your three columns show slightly different bullet counts at three prices, a rational buyer optimises the one dimension they trust, which is cost per seat. You built that result. If the columns differ by how much operational risk moves from the client to you, the same rational buyer starts asking which risks they are currently carrying alone. That is a conversation a weaker provider cannot follow you into.
Section 2
Why three tiers, and what breaks at two or at five
Three is usually right, and the reasons are structural rather than aesthetic. Two tiers reads as a real package and a crippled one. The buyer assumes the cheap column is a trap and the expensive column is the actual price, so the sheet does no sorting work. Five or more tiers fails from the other end. Every extra column is another surface where a competitor's sheet can look better on one line. Sales calls turn into configuration sessions, your delivery team ends up supporting five different definitions of standard, and buyers who cannot judge technical quality respond to a long menu by deferring the decision or defaulting to the cheapest thing. Three gives you the minimum structure that does three separate jobs. The bottom tier sets a floor and disqualifies the wrong buyer, which is its main function. It should be a real service you are willing to deliver, not a decoy, and it should be priced so that a business unwilling to pay for the floor removes itself early. The middle tier is the sale you intend to make. The top tier catches the regulated or high-risk client and, as a side effect, makes the middle look proportionate. The cost of three is honest and worth stating. You have to be able to deliver the top tier. A top tier you have never actually run is a liability the first time somebody buys it, and someone eventually will. If you cannot staff an annual tabletop exercise or produce compliance evidence on demand, do not print it. An undeliverable promise is a negative signal the moment it is tested, and in a referral market a failed delivery travels further than a won deal.
Section 3
The ladder rule: tiers differ by risk transferred, not by feature count
The common mistake is building tiers by addition. Tier 2 is tier 1 plus two more bullets. Tier 3 is tier 2 plus more support. That is not a ladder, it is a bigger number with decoration. A defensible ladder moves risk from the client to you as the price rises. Concretely, one shape that holds up: Foundation. Monitoring and alerting, patch management on a stated cadence, endpoint protection, backup with a named retention period, help desk during defined business hours, a quarterly report. Response targets stated, no credit attached. Managed, which is your intended sale. Everything above, plus after-hours coverage, identity and access management, restores tested at a stated frequency rather than backups merely running, vendor liaison so the client stops chasing their line-of-business software provider, a named engineer, and a quarterly technology review with a 12-month budget forecast. A response commitment with a service credit behind it. Governed. Everything above, plus documented alignment to a security framework, an annual incident tabletop, log retention for a stated period, a named virtual CIO with a written roadmap, evidence packs assembled for auditors and insurers, and priority sequencing during a mass incident. Notice what climbs the ladder: the items that are expensive for you to deliver and impossible for a disorganised firm to fake. Tested restores, evidence packs and mass-incident priority all cost real labour and require real process. That is precisely why they work as signals. Bullets that cost nothing to promise belong at the bottom or nowhere. The practical test on any line item you are about to print: could a two-person shop with no documentation truthfully offer this tomorrow? If yes, it is not differentiating your tier, it is padding it.
Section 4
Scope boundaries: three lists, and the discipline to enforce them
Scope creep is not a client behaviour problem. It is a document problem that becomes a behaviour problem. Every tier needs three lists, not one. What is included without limit. What is included up to a stated limit. What is out of scope and billed separately. For example: unlimited remote help desk for supported applications; onsite work included up to a set number of hours per month; project work, cabling, hardware procurement and third-party application development explicitly outside the agreement. Then define the words the lists depend on, because these are where disputes live. What is a supported device, and what happens to an unsupported one. What is a supported application. What are business hours, stated with the timezone. And the pair that causes the most damage when left vague: response and resolution. Response means you have acknowledged and assigned. Resolution means it is fixed. A response commitment you can keep is worth more than a resolution promise you cannot, and mixing the two into one line about fast support gives the client a reasonable case for expecting the second when you meant the first. The uncomfortable part is enforcement. A boundary is only real if somebody says no out loud. If the owner overrides the limit every time a client pushes, the tier sheet is decorative and the team learns that the document does not govern anything. Set an escalation rule: out-of-scope requests get quoted, not absorbed, and the person quoting is not the person on the call taking the pressure. If you want the adjacent pieces on how this connects to pricing and pipeline in this market, we keep them together at https://bizgrowthaxel.com/msp/.
Section 5
Publish the list of what moves the price
Buyers do not fear a high price nearly as much as they fear an unpredictable one. The variable that keeps a prospect from signing is often not your rate, it is the suspicion that the rate is the opening figure and the invoices will drift upward for reasons never explained. So publish the multipliers. A short list, stated plainly, of what raises the number and why: Seat count and device count, because both drive support labour. Number of physical sites. Servers, and whether any of them sit on premise. After-hours or weekend coverage requirements. Regulated data, which adds documentation and audit obligations rather than just technical work. Legacy or end-of-support systems still in production, which raise both labour and your own exposure. The size of the onboarding backlog, meaning how much undocumented mess you inherit in month one. And whether the client wants you to own their vendor relationships rather than only their infrastructure. Each of those lines raises price because each raises your labour or your risk. Say that explicitly. It converts a negotiation into a qualification, which is the outcome you want. A prospect running three unsupported servers now knows before the call that they are not a floor-tier client, and either accepts that or removes themselves. It also gives you something to hold up when a competitor quotes 30 percent under you on seats alone. You are not arguing that you are better, which is unverifiable. You are pointing at the lines their quote does not price, which is checkable. That shifts the burden of explanation onto them.
Section 6
Anchors, decoys, and why the middle tier gets chosen
Two well-documented effects do most of the work in a three-column layout, and both are worth understanding before you arrange the columns. Anchoring means the first number a person sees sets the frame for every number after it. If you lead with your cheapest tier, every other tier reads as an increase over a reference point you supplied. Lead with the governed tier and the middle reads as a reduction. Same prices, different felt magnitude. The second is the compromise effect. When people cannot judge quality directly, they avoid extremes. They assume the cheapest option is missing something they will regret and the most expensive is more than they need, so they take the middle. That is not a quirk to exploit, it is a rational rule of thumb under genuine uncertainty, which is exactly the position your buyer is in. The design implication is simple: the middle tier must be the package you actually want to deliver, at a margin that survives a heavy-usage client. If your intended sale is your thinnest-margin package, the mechanism is running against you every day. There is a line here worth being clear about. A decoy tier built so nobody can sensibly buy it is a trick, and experienced buyers find it in the first call. A top tier that a regulated client genuinely needs performs the same anchoring function and survives scrutiny, because it is real. The difference costs you nothing to get right and everything to get wrong, since the entire premise of packaging is that you are sending signals a weak competitor cannot send. A fake tier is a cheap signal wearing an expensive costume.
Section 7
Where this ages, and what the model cannot see
Treat every package as dated. What counts as standard scope is set outside your business, and two shifts can reset it. The first is vendor consolidation. When a security or backup capability gets absorbed into a platform licence your client already pays for, a line that differentiated your middle tier last year becomes a checkbox this year. You keep charging for it, a competitor stops, and the comparison quietly turns against you without anyone telling you why. The second is an AI shift in tier-1 support. If a meaningful share of routine tickets stops reaching a human, per-seat pricing tied to headcount detaches from the labour it was proxying for. A competitor operating on the new cost base can price below you while earning the same margin, and you will read it as irrational discounting rather than a changed cost structure. Both age a fixed package, which is why a twice-yearly review is the minimum. The question to ask each time: which line item was a differentiator 12 months ago and is now simply assumed? Now the limit of this entire way of thinking. It assumes a comparison is happening. In referral-driven deals, which is most of the pipeline for most MSPs under 50 people, there frequently is no bake-off. The prospect called because a peer vouched for you, and the decision was substantially made before your tier sheet was opened. Signaling and mechanism design are silent on that. They cannot see who owes whom a favour, whether the referrer's own credibility is on the line, or whether the incumbent provider's owner sits on the same board as the CFO. No packaging change moves any of that. So be honest about what better packaging buys you in a referral-heavy business. It improves margin, it prevents scope creep, and it makes delivery sane. It will not obviously raise your close rate on deals you were already going to win.
Section 8
Fitness test: are you ready to repackage
You are ready for this if most of the following is true. You have delivered enough contracts to know what your actual cost to serve looks like at different client shapes, rather than guessing. You can staff every commitment in your top tier today, including the ones nobody has bought yet. You have someone other than the owner who can say no to an out-of-scope request and make it stick. You are losing deals on price against firms you believe are worse, which suggests the comparison frame is the problem rather than the price. And you have a channel where strangers evaluate you before speaking to you, because that is where a comparison frame does its work. You are not ready if any of these hold. Your pipeline is entirely referral and your close rate is already high, in which case repackaging is a margin and delivery project, not an acquisition one, and you should scope it that way. You do not yet know your gross margin per client, because you will price the middle tier by feel and discover the error at scale. You are still delivering ad hoc break-fix work under a monthly label, since packaging that honestly means naming what you do not do, and you may not be willing to. Or you are considering a top tier you cannot deliver in order to make the middle look better, which will cost you more than the pricing lift is worth. If you are in the not-ready column, the first move is not a new tier sheet. It is measuring cost to serve on your existing clients for one quarter. Everything above depends on knowing which promises you can afford to make.