Company Accounts, Buyer Roles, and Approval Flows: A Playbook for B2B Self-Service

A buyer with a large order and no way to route it past their own finance lead will call your competitor, or worse, will call your sales rep and become a manual order again. Self-service is not the catalog. It is the permission and approval logic underneath the catalog, and there are two separate approval systems involved. Most merchants build the one that protects their own business and never build the one their buyers actually need.

The first system approves a company to buy from you: credit, terms, catalog access. The second routes a purchase through the buyer’s own internal chain before it becomes an order. Both are necessary. Only the first appears in most implementations, and the second is where large orders quietly stall.

This is a design playbook rather than a feature list. Work through it and you will have a role matrix, an approval pattern, and a defensible answer to the question of what your buyers can do without contacting you.

Two approval systems, one of which is usually missing

Merchant-side approval answers whether this company can buy from you at all, and on what basis. It covers account application and vetting, credit assessment and payment terms, which catalog and price list the company sees, and which locations or entities are recognised. This is well understood, it is where platform features concentrate, and it is genuinely important.

Buyer-side approval answers whether this person, inside that company, can commit this amount of their employer’s money. A procurement assistant assembles a two hundred line order and needs it authorised by a department head. A store manager reorders within a limit and needs nothing. A regional buyer can order for three locations but not a fourth. None of this is your risk. It is your buyer’s governance, and if your store cannot express it, the buyer resolves it outside your store, which means email, spreadsheets, and eventually a phone call.

The consequence is measurable but invisible in a standard funnel. An order that stalls waiting for internal authorisation leaves no failed checkout and no abandoned cart event worth acting on. It looks like an account that went quiet. Our analysis of the B2B conversion levers a DTC audit never inspects covers why this class of loss sits entirely outside a session-based funnel.

Which system is costing you: three questions

Answer these before designing anything.

1. How many orders still arrive by email or phone from accounts that already have a login? If the answer is more than a handful, ask what those buyers could not do themselves. The answers cluster: could not see their price, could not add a colleague, could not get a large order authorised, could not reorder without rebuilding the cart.
2. What is the value distribution of your self-service orders? If online orders cluster below a threshold while large orders arrive manually, you are looking at a buyer-side approval gap. Buyers are self-serving what they can authorise alone and routing everything else around you.
3. How long does a new company account take from application to first order? If it is more than a few days, the merchant-side process is the constraint, and the fix is procedural before it is technical.

Question two is the one most merchants have never asked, and the value distribution usually answers it in a single chart.

Branch A: merchant-side onboarding

Four decisions, and the goal is to make them explicit rather than to make them permissive.

What triggers approval. Decide the minimum information that lets you approve an account, and separate it from information you would like to have. A long application form is the most common reason a wholesale account never completes signup.

Who approves, and against what. A named owner with written criteria, not a judgement call by whoever opens the inbox. Credit checks and trade references belong here, and so does a clear rule for the borderline case.

What an unapproved buyer sees. A gated experience, a public catalog without prices, or a limited catalog. Each is defensible; what causes damage is having no deliberate answer, so the buyer sees retail prices and draws conclusions.

What happens on approval. Which catalog, which price list, which payment terms, which locations. This is the point where a company record becomes a commercial relationship, and it should be a configuration rather than a series of manual steps.

Time to first order is the metric worth watching here. Every day between application and a buyer’s ability to place an order is a day they are ordering from someone else.

Branch B: buyer-side roles and routing

This is the branch that is usually absent, and it starts with a role matrix rather than with a workflow.

Most B2B platforms provide company records with locations underneath them and a small set of roles per user. On Shopify, that means companies, company locations, and role assignments that determine who can order for which location. Deeper conditional logic, such as value thresholds routing to a specific approver, is generally built with additional configuration or apps rather than provided natively, so it belongs in a scoping conversation rather than being assumed.

Design the matrix before choosing mechanisms. Four dimensions cover most real buyer organisations:

A workable default for mid-market buyers is three roles. An ordering user with one location, the entitled catalog, and a value ceiling. A location admin with one location, no ceiling, and the ability to add users and view invoices. And an account owner with all locations, no ceiling, and full administration. Start there and add complexity only where a real customer requires it.

The trap here is designing the matrix around your largest customer. Enterprise buyers have genuinely complex governance; most of your accounts have two people and no approval chain at all, and a system built for the former makes the latter unusable.

Four approval patterns, and which to build

Two observations from the table. Threshold approval is what most buyers describe when they ask for approvals, and it is the one most likely to need custom work. Cart handoff is frequently underrated: it solves the same problem with a fraction of the build, because it moves the governance into the buyer’s hands rather than modelling their policy in your system.

That last point generalises into the design principle worth taking from this article. You are not trying to reproduce your buyer’s approval policy. You are trying to give them enough structure to apply it without leaving your store. The former is unbounded work that changes whenever your customer reorganises. The latter is finite.

The default sequence

Five steps, in the order that produces working self-service fastest.

1. Fix time to first order. Reduce the application to what you genuinely need, name the approver, write the criteria. This is procedural work and it usually unlocks more revenue than anything else on the list.
2. Get entitlement right. Every logged-in buyer sees their correct catalog and their correct price, every time. A buyer who sees retail pricing once frequently stops trusting the store and reverts to email without telling you.
3. Build the three-role matrix. Scope, catalog, value ceiling, administration. Resist adding a fourth and fifth role until a real customer needs one.
4. Add cart handoff. Low cost, and it removes the most common reason large orders leave the store.
5. Add threshold approval only where accounts require it. Scope it against named customers with a stated policy, since threshold logic built speculatively tends to model a policy nobody actually has.

Steps one and two are prerequisites rather than a phase. Approval logic layered on top of broken entitlement produces a system where buyers are routed carefully toward the wrong price.

Flatline builds B2B self-service across Shopify Plus and other stacks, and the recurring finding is that the role matrix is a commercial document rather than a technical one. Somebody in sales knows how each account actually buys, and that knowledge is usually not written down anywhere.

What sits next to this

Two adjacent decisions are deliberately out of scope here and each has its own treatment. Pricing governance across wholesale and retail decides whether your DTC pricing undermines the partners who stock you, which is a partner trust question as much as a margin one. And the wider operating cost of running both channels, covered in where the hidden operating cost lives, is what the approval design sits inside.

For the platform-level feature picture underneath all of this, our overview of Shopify B2B features arriving on every plan covers what is native and what is not.

Frequently Asked Questions

What is the difference between merchant-side and buyer-side approval? 

Merchant-side approval decides whether a company can buy from you and on what terms: vetting, credit, catalog, and price list. Buyer-side approval routes a purchase through the buyer’s own internal authorisation before it becomes an order. Most implementations build the first and omit the second, which is why large orders leave the store and arrive by email instead.

How many buyer roles do we actually need? 

Three covers most mid-market accounts: an ordering user with a value ceiling, a location admin with authority for one location, and an account owner with full scope. Add roles only when a named customer’s governance genuinely requires it. Designing the matrix around your most complex enterprise account makes the system unusable for the accounts that make up most of your revenue.

Does approval routing need to be built, or is it native? 

Company accounts, locations, and basic role assignment are generally native on B2B-capable platforms. Conditional routing, such as value thresholds sending an order to a specific approver, usually requires configuration or an app. Confirm what your platform provides before scoping, and consider cart handoff first, since it solves much of the same problem for far less build.

Why do large orders stall even when the account is set up correctly? 

Because the buyer cannot complete the purchase alone and has no way to route it internally from inside your store. The order does not fail visibly; it simply never gets placed, and the account looks quiet. Check the value distribution of your self-service orders: a ceiling in the data with larger orders arriving manually is the signature of a missing buyer-side approval path.

Key Takeaways

• Two approval systems exist and most merchants build one. Merchant-side approval protects your business; buyer-side routing lets a purchase clear the buyer’s own governance without leaving your store.
• Diagnose with the value distribution of self-service orders. A ceiling in that data, with larger orders arriving by email, is the clearest signal of a missing approval path.
• Design a role matrix before choosing mechanisms. Scope, catalog, value ceiling, and administration cover most real buyer organisations, and three roles are enough to start.
• Do not model your buyer’s approval policy. Give them enough structure to apply their own, which is finite work, rather than reproducing a policy that changes whenever they reorganise.

Self-service succeeds when a buyer can complete a normal purchase without contacting anyone, including their own colleagues, for anything routine. Everything above is in service of that one sentence, and the value distribution of your own orders will tell you how close you currently are