Mark
A DAO marketing grant can fund useful education and help people understand a protocol. It can also consume treasury assets without showing whether anyone learned about or used the product. The difference is not always obvious from a polished proposal or a large promised audience.
Before voting, members can examine the request as a series of testable claims: who the campaign should reach, what work will be delivered, how results will be measured and who can authorise each payment. This process is useful even for token holders who do not consider themselves marketing specialists.
Start with the campaign brief, not promises about token price
A usable brief identifies the audience, the problem the campaign addresses, the proposed channels and the action the team wants people to take. For example, an education campaign might aim to bring prospective users to a wallet-safety guide, rather than simply promise to make the protocol trend. It should explain why that audience and channel are relevant to the DAO.
Reviewers who have appropriate campaign reports and creative briefs can use a workspace such as OhlasAi to help analyse the supplied material and review creative work against the stated goal and audience. Its output is a prompt for further questioning, not independent evidence that the proposal is sound, that an audience consists of genuine users or that a token will perform in a particular way. Reviewers still need to check factual claims and decide what information is appropriate to upload.
Separate a marketing objective from a market outcome. A team can propose publishing tutorials, improving a landing page or testing paid messages; it cannot credibly guarantee a token-price rise. Prices are affected by many forces outside a campaign, including liquidity, market conditions and holder behaviour. Even a rise during the campaign would not establish that the marketing caused it.
Check that the requested activity fits the DAO’s actual priorities. A developer-focused protocol may gain less from a broad consumer awareness campaign than from clear technical documentation or targeted workshops. Ask the proposer to state what the DAO would learn if the campaign fails, as well as what a useful result would look like. If the underlying product is not ready for the audience, spending more on promotion may simply send people to a poor experience.
Check whether the evidence supports the requested grant
DAOs differ in how decisions are proposed and approved, but governance generally gives members a way to coordinate decisions over shared resources. Ethereum.org’s What is a DAO? explains proposals, voting and collective treasury decisions. A vote, however, is only as informed as the information available to voters; a popular proposal is not automatically a well-supported one.
Start with the baseline. If a proposal claims that it will increase qualified website visits, ask for the current number, the period measured, the source and the definition of “qualified.” If the DAO has no reliable baseline, an initial measurement phase may be more sensible than a large performance-based budget. A comparison with a previous campaign is only useful when the audience, timeframe, channel and measurement method are sufficiently similar.
Inspect the evidence behind reach and engagement claims. An account’s follower count is not the same as unique people reached, and impressions do not establish that someone understood a message or used a product. Ask for source reports where available, the dates covered, paid versus organic distribution and an explanation of unusual spikes. Screenshots can support a claim, but they rarely show enough context on their own.
Be especially cautious with on-chain activity. A wallet address is not necessarily a unique person: one person can control several addresses, and an automated system can generate transactions. Ask how the team will identify the relevant action, filter obvious internal or test activity and handle incentives that may attract short-term activity rather than lasting use. A campaign link or referral code can help trace a path, but it does not prove that marketing was the sole cause of a transaction.
Look for alternative explanations. A product launch, token incentive or broader market rally occurring at the same time could affect traffic and activity. Where practical, compare results with a similar period or an unpromoted audience, while acknowledging differences between the groups. The proposal should disclose uncertainties and include disappointing results, not just select the most flattering metric after the fact.
Finally, consider the provenance of the materials. Who produced the previous report, who had access to the underlying data and can another reviewer reproduce the calculation? The DAO should avoid publishing private user information just to demonstrate a result. Aggregated figures, controlled access and a clear retention policy can make reporting more useful without exposing unnecessary data.
Turn broad ambitions into deliverables and funding milestones
“Grow the community” is an ambition, not a deliverable. A stronger proposal specifies the work: for example, a set of educational articles, a landing-page revision and a report on a defined advertising test. For each item, identify the owner, due date, format and place where members can inspect it. Agree on review criteria before work begins, including who may request revisions and how disputes will be handled.
Separate outputs from outcomes. An output is something the contractor can submit, such as a published guide or a campaign report. An outcome is the effect the DAO hopes to see, such as more readers completing a tutorial. Both matter, but a contractor may not control the outcome. Avoid a payment structure that rewards only a convenient activity count while ignoring quality, or that demands an outcome no one can reliably attribute to the work.
One possible structure is to fund a small discovery phase, review its brief and measurement plan, then approve a limited pilot before releasing a larger production budget. A final stage could cover delivery and a report comparing the agreed metrics with the baseline. This is an illustration, not a universal allocation: the right amounts and intervals depend on the DAO’s rules, the work and the risk of paying in advance.
For every release, specify the evidence required and the decision maker. A milestone might require links to completed content, an itemised invoice and access to the underlying campaign report. It should also state what happens if the work is late or incomplete: whether the recipient can revise it, whether unused funds must be returned under the agreement and who resolves disagreements. Do not assume that a governance vote alone creates an effective refund mechanism.
Execution controls are separate from performance review. A multisignature wallet can require a threshold of authorised signatures before a transaction is executed. The documentation for Running the Safe Transaction Service describes collecting transaction signatures toward that threshold. Signer approval does not independently prove that a marketing milestone was achieved; the DAO still needs an agreed process for checking evidence before signers approve a payment.
Scrutinise software costs, contractor fees and treasury exposure
A budget should show quantities and assumptions, not just a single total. Break out contractor time, creative production, advertising spend, software, reporting and any contingency. Ask which costs are fixed, which depend on usage and whether the recipient has discretion to move money between categories. If a line item is uncertain, a capped pilot and a later request can be more informative than approving a large estimate upfront.
Software pricing is one example of a checkable assumption. At https://www.ohlas.uk/pricing, OhlasAi lists a €10 minimum purchase, a rate of €1 per 10,000 credits and one-time payments without automatic renewal. A proposal using it should still estimate the intended tasks and expected credit consumption, then explain how that estimate was obtained. Credits purchased are not tasks completed: the published conversion rate does not establish a cost per audit or prove that the tool is cost-effective for the DAO.
Apply the same scrutiny to people and distribution. For contractor fees, ask for the role, expected hours or deliverables, rate, review responsibilities and relevant experience. For advertising, distinguish the fee paid to manage a campaign from the amount actually spent on placements. If the proposal includes influencers or partners, identify the proposed compensation, required disclosures and evidence that their audience matches the campaign.
Consider treasury denomination as well as the headline amount. A grant denominated in a volatile token can change substantially in value between the vote and payment; a grant denominated in a stablecoin has different asset and counterparty risks. The proposal should say which asset will be paid, whether the amount is fixed in that asset or converted at a specified time, and who bears conversion costs. It should not quietly treat a forecast token price as a guaranteed budget.
Check the recipient address and payment permissions through the DAO’s established process, especially if a new address is supplied shortly before execution. Confirm who controls the receiving wallet, whether onward payments are permitted and how remaining funds will be accounted for. Never ask a grantee to share a private key or seed phrase as proof of identity. For complex arrangements, members may need qualified legal or accounting advice rather than relying on a marketing proposal to settle compliance questions.
Use a pre-vote checklist and define when funding should stop
Before voting, a member should be able to answer a few questions from the proposal and its supporting documents. If the answers are missing, request revisions rather than filling the gaps with assumptions:
- Purpose: Who is the audience, what action is sought and why does it support a stated DAO priority?
- Evidence: What is the baseline, where did it come from and what limitations affect the proposed measures?
- Work: What will be delivered, by whom, by when and against which acceptance criteria?
- Money: How are costs calculated, which asset will be paid and what spending flexibility is allowed?
- Control: Who reviews each milestone, who can authorise payment and what happens to unused or disputed funds?
- Reporting: When will results be published, which sources can reviewers inspect and how will user privacy be protected?
Agree on pause conditions before the grant starts. Missing reports, material budget changes, unapproved subcontracting or evidence that a campaign is reaching the wrong audience may justify stopping the next payment while the DAO reviews the situation. A weak pilot result need not mean misconduct: the agreed response might be to revise the message, end the experiment or decline to scale it. The important point is that the next decision should not be automatic.
After the work ends, compare spending and results with the original plan and record what remains uncertain. If the campaign met its publication obligations but did not produce a measurable change in use, say both. If the data is too weak to reach a conclusion, say that too. A good grant review protects the treasury not by demanding certainty, but by making the cost, evidence and next decision visible to the people who govern it.
