The Market Committee Vote That Decides Everything: 9 Governance Questions to Settle Before a Farmers Market Sells Online

Why farmers market online launches stall on governance, not software

Online farmers market launches stall because configuration questions get escalated to a committee that only meets monthly. Commission rate, refund liability, who owns the customer list: every one is a settings screen in the software and a governance decision in the constitution. The software waits for the vote.

The pattern is consistent. A market signs up, an administrator starts configuring, and within the first afternoon hits a field that nobody is authorised to fill in. Does the market take 8% of an online order or charge a flat online stall fee? Is the refund the stallholder's cost or the market's? The administrator sends an email, the chair says it needs to go to the committee, and the committee meets in three weeks. Repeat that four times and a six-week launch becomes a five-month one.

None of these are technology problems. They are questions about what the market actually is once it has a website: a landlord renting digital space, or a retailer that buys and resells. Markets have run for decades without answering that, because at a physical stall the answer is obvious. Cash goes from shopper to stallholder and the market collects rent. The moment one checkout collects money for forty stallholders, the ambiguity becomes a legal and accounting fact that somebody has to decide.

Decisions 1 to 3: money, seller of record, and who holds the merchant account

These three decide the market's legal identity online and must be settled together. Whether the market or the producer is seller of record for VAT, who holds the merchant account that receives card payments, and whether the market takes commission or a flat online fee are one interlocking question, not three independent ones.

Two coherent models exist. In the agent model, each stallholder remains seller of record, the market acts as a disclosed agent, the customer's invoice names the producer, and the market charges a fee for the service. This keeps small producers under their own VAT registration thresholds and keeps the market's turnover small, but it needs software that can issue per-producer invoices from one basket and split one card payment across many payees. In the merchant model, the market buys from producers and sells to the shopper. Invoicing is simple and the market controls pricing and refunds, but the market's declared turnover now includes the full retail value of everything sold and the market carries the tax registration and the product liability that comes with being the retailer.

The merchant account follows from that choice and cannot be decided separately. If the market is seller of record, it holds the account and settles producers by payout run. If producers are sellers of record, the platform needs split payments or a clearly documented client-account arrangement where the market holds money on behalf of producers. Markets that chose the agent model for tax reasons and then opened a single ordinary bank account in the market's own name created a mismatch that their accountant found at the first year end. Decide the tax position first, then open the account that matches it.

Motion 1 (seller of record): "That for online sales, the [market / individual stallholder] is the seller of record, with the market acting as disclosed agent where applicable, and that customer invoices are issued in the name of the [market / stallholder]."
Motion 2 (merchant account): "That online card payments are received into an account held by [the market / each stallholder via split payment], and that funds held on behalf of stallholders are recorded separately from market funds."
Motion 3 (market revenue): "That the market's revenue from online sales is [X% commission on the value of goods sold / a flat fee of €X per stallholder per month / a per-order fee of €X], reviewed after [6/12] months."

Decisions 4 to 6: whether a stallholder can refuse an order, price parity, and non-participation

These three decide what an online listing promises the shopper. Whether a stallholder may decline a confirmed online order, whether online price must match stall price, and what happens to stallholders who do not join determine whether the storefront is a guarantee or a request, and whether the market can advertise itself as the market at all.

Refusal rights are the one most often skipped and the most expensive to skip. A grower who declared ten kilos of courgettes on Tuesday and lost half the crop on Wednesday will not fulfil ten kilos on Saturday, so the question is never whether refusals happen but what the customer experience is when they do. Markets generally land in one of three places: no refusal after cut-off, with the market sourcing a substitute from another stallholder; refusal permitted up to a stated hour with an automatic refund of that line; or full substitution rights where the picker swaps a comparable item. Each one is a different notification flow, a different refund rule and a different set of permissions in the producer app, which is why the answer has to exist before the configuration does.

Price parity and non-participation shape the catalogue. If online price may exceed stall price to cover the market's cut, shoppers standing at the stall with their phone will notice, and someone has to be able to explain why. If parity is mandatory, the market's fee has to come out of the producer's margin, and the producers need to be told that in the same meeting. Non-participation is separate again: a market with 40 stalls and 12 online sellers is still a market, but the website copy, the search categories and the pickup messaging all have to be honest about it, and the committee should say explicitly whether joining is voluntary, opt-out by default, or a condition of the stall licence at the next renewal.

Motion 4 (refusal): "That a stallholder may decline an online order line up to [time] on [day], after which the order is binding; declined lines are [refunded automatically / substituted by the market with a comparable item at no additional cost to the customer]."
Motion 5 (price parity): "That online prices [must match / may exceed by no more than X% / are set freely by] the stallholder's market-day price, and that the market's fee is borne by the [stallholder / customer]."
Motion 6 (participation): "That online participation is [voluntary / opt-out / a condition of stall licence from DATE], and that non-participating stallholders [are / are not] listed on the website as market-day only."

Decisions 7 to 9: who owns the customer, who answers the email, and who eats the shortfall

These three decide who carries the operational load after launch. Customer ownership determines what a stallholder may see and do with shopper data, the support question decides whose inbox and phone number appear at checkout, and the shortfall rule decides who pays when goods ordered do not arrive.

Customer ownership is a GDPR question dressed as a commercial one. If the market is controller of the shopper record, stallholders see order lines and pickup names but never the full contact list, and the market can run its own newsletter. If stallholders are joint controllers of their own customers, they can market directly, and the market has to write that into a data sharing agreement and a privacy notice the shopper actually sees. Markets that leave this unvoted usually discover it when one popular grower asks for a spreadsheet export of everyone who bought their cheese, and there is no policy to point to.

Support and shortfall are the ones that decide whether the launch survives month three. Someone has to own the checkout email address, answer within a stated window, and have authority to issue a refund without convening anyone. That is a named role with stated hours, not a shared inbox. Shortfall is the same question on the money side: when a grower delivers six kilos against a ten kilo order, the customer is refunded four kilos of value, and the committee has to say whose money that is. The workable options are that the producer's payout is reduced by the undelivered value, that the market absorbs it from commission as a service cost, or that a tolerance band applies below which nobody is charged. Choose one, write it into the stallholder agreement, and the reconciliation run has an answer every week instead of an argument.

Motion 7 (customer ownership): "That the market is data controller for shopper accounts; stallholders receive only the order data necessary for fulfilment and may not export or use shopper contact details for their own marketing."
Motion 8 (customer support): "That [role] is responsible for all online customer contact, responding within [X] working hours, with authority to issue refunds up to €[X] without further approval."
Motion 9 (shortfall): "That where a stallholder delivers less than the ordered quantity, the customer is refunded in full for the shortfall and the [stallholder's payout is reduced accordingly / market absorbs the amount from its commission], subject to a tolerance of [X]% on weight-variable items."

Running the vote: one agenda, one meeting, nine motions

Put all nine motions on a single agenda and vote them in one sitting before any vendor demo is booked. Decisions 1 to 3 must be taken together because they are legally interdependent; decisions 4 to 9 can be voted individually. Record each outcome as configuration input, not as minutes prose.

The practical sequence is to circulate a one-page paper a week ahead that states each question, the two or three real options and the downstream consequence of each, then hold a session with the accountant available for the tax question. Invite three or four stallholders who will be affected, because motions 4, 5 and 9 change how they get paid and a decision imposed without them in the room gets re-litigated at the first shortfall. Expect the refusal and shortfall motions to take the longest; they are where the market discovers what it actually believes about risk.

Write the outcomes into a short online trading annexe to the stall licence rather than into the constitution. The annexe should carry a review date, typically six or twelve months, because at least two of the nine will turn out to be wrong. Commission rates get revised once real order values are visible, and refusal windows get tightened once stallholders see how a Friday decline lands on a Saturday picker. A decision with a review date is much easier to pass than one that feels permanent.

Circulate the nine questions with options and consequences one week before the meeting.
Have the accountant or bookkeeper present for motions 1 to 3.
Include affected stallholders for motions 4, 5, 6 and 9.
Vote 1 to 3 as a single resolution; vote the rest individually.
Record outcomes in a dated online trading annexe with a review date.
Only then book vendor demos, carrying the nine answers as your requirements list.

What each answer becomes once you configure the software

Every one of the nine answers maps to a specific setting, permission or document in a multi-producer platform. Bring the decisions to the vendor as requirements and configuration takes days. Bring them as open questions and the vendor picks defaults for you, which is how markets end up with a commission model their constitution does not support.

The mapping is direct. Seller of record determines whose name prints on the invoice and whether one basket generates one invoice or several. Merchant account choice determines whether the platform runs split payments or a payout run. Commission versus flat fee determines the settlement calculation and what appears on a producer's remittance. Refusal rights determine whether the producer app has a decline button and what the cut-off enforcement does. Price parity determines whether producers hold one price list or two. Customer ownership determines what a producer sees on an order line: a name and a pickup slot, or a full contact record. Support ownership determines which address sends transactional email. Shortfall policy determines the reconciliation rule that runs every week without human intervention.

Some platform defaults exist precisely because a market chose badly and had to unpick it. A default that hides shopper contact details from producers, a default that requires a declared cut-off before any storefront goes live, a default that reduces payout by undelivered value unless overridden: each of those is the safe position when governance is unclear, and each can be changed once the committee has actually voted. At Plodie we would rather configure a market than correct one, which is why the first conversation with an operator tends to be about these nine questions rather than about features.

Key Takeaways

Online market launches slip on unresolved committee questions, not on software; each unanswered question is a configuration field nobody is authorised to fill in.
Seller of record, merchant account and market revenue model are legally interdependent and must be voted as one resolution, with the accountant in the room.
Refusal rights, price parity and participation rules decide what an online listing promises the shopper, and each maps to a different notification and refund flow.
Customer ownership, named support responsibility and a written shortfall rule decide whether the launch survives past month three.
Record all nine outcomes in a dated online trading annexe to the stall licence with a six or twelve month review, then book vendor demos with the answers as your requirements list.

If your committee lands on split payments and per-producer invoicing, it is worth reading why Shopify breaks at the second producer, which covers the six points where a single-merchant storefront fails a multi-producer market.

Frequently Asked Questions

Does a farmers market need to register for VAT to sell online?

It depends entirely on whether the market is seller of record. If the market buys from stallholders and resells, the full retail value counts toward the market's turnover and registration thresholds apply. If the market acts as a disclosed agent and each stallholder remains seller of record, only the market's commission or fee counts as its own turnover. Confirm the position with an accountant before choosing a payment setup, because the two models need different bank arrangements.

How long does it take to launch an online farmers market once the decisions are made?

With the governance settled, the limiting factors are catalogue building and producer onboarding rather than configuration. Expect a few weeks to photograph and list products, train stallholders to declare availability, and run a parallel pilot with a small group of shoppers before opening publicly. Markets that arrive with open governance questions typically add months because each unanswered question waits for the next committee meeting.

What if only a few stallholders want to sell online?

That is workable and common at the start. A market with 40 stalls and 10 online sellers can launch, provided the website is honest about which producers are available online and the pickup messaging does not imply the whole market is represented. Voluntary participation with a review date usually gets more sign-ups over time than making it a licence condition on day one, because reluctant stallholders get to watch the first cohort get paid.

Should a farmers market charge commission or a flat monthly fee for online sales?

Commission scales with volume and keeps the market's interest aligned with producer sales, but it makes small orders relatively expensive to process and it complicates price parity. A flat fee is predictable for stallholders and simple to invoice, but it can mean the market runs the storefront at a loss in the early low-volume months. Many markets start with a flat fee during the pilot period and switch to commission once order volume is visible, which is why the motion should carry a review date.

Who is liable if a customer gets ill from something bought through the market's website?

Liability tends to follow the seller of record. If the market is the retailer, it carries retailer obligations including product liability and recall duties; if the stallholder is seller of record with the market as agent, primary responsibility stays with the producer. Either way the market needs traceability records linking each order line to a producer and a batch, and its insurer should be told that the market now sells online.