Wallet installs are the easiest metric in Web3 to move. Run a quest campaign, promise a reward, watch the number climb.
Then look at how many of those wallets ever received a single transfer.
The gap between install and first deposit is where the entire product lives, and it’s the part that install-focused campaigns never touch.
Why the funnel breaks
Installing is free. Funding is a decision with real consequences, and the user hits several walls between the two.
The on-ramp rejects them. Wrong country, wrong card, KYC that fails, a provider that doesn’t cover their bank. This is the single largest drop-off in most wallet funnels and it’s largely invisible unless you instrument the on-ramp specifically.
The seed phrase screen scares them. Twelve words, a warning about irreversible loss, and a checkbox. For a first-time user this is genuinely alarming, and a meaningful share simply close the app.
They need gas they don’t have. The user funded with a token and now can’t move it because they have no native asset for fees. This is a solved problem — sponsored transactions, account abstraction, gas relayers — and plenty of wallets still ship without solving it.
They had one reason and it ended. They downloaded the wallet to claim something. They claimed it. There was never a second use case.
What to measure
Funded, retained wallets. Accounts that received value and were still active thirty days later.
Everything upstream is diagnostic:
- Install to account creation
- Account creation to first funding attempt
- Funding attempt to funding success — instrument the failures, this is where the money is
- First funding to first useful transaction
- 30-day and 90-day retention
A wallet with two million installs and forty thousand funded accounts doesn’t have an acquisition problem. Buying more installs makes the ratio worse and the report look better, which is a bad combination.
Channels that actually work
Distribution partnerships. Being the recommended wallet inside a protocol, a game or an exchange your target users already use. Highest quality of any channel, and the hardest to buy — it’s a business development effort, not a media buy.
Search on high-intent terms. People looking for a wallet that supports a specific chain, or that does a specific thing, convert far better than people responding to an ad.
Creator content in target regions. Wallet choice is heavily influenced by local trust. A creator in Vietnam or Nigeria walking through the app in the local language does something no English-language campaign replicates.
App store optimisation. Unglamorous, consistently underrated, and cheap. Screenshots, description, keyword coverage and review volume all move install-to-open rates.
Broad paid install campaigns. The worst quality in this category, because they select for people responding to an ad rather than people with a need. Useful for volume tests, poor as a core strategy.
The onboarding decisions that matter most
Lead with the job, not the philosophy. People download wallets to buy something, join something or claim something. Start there. The self-custody conversation lands much better once the user has a reason to care.
Offer recovery that doesn’t break the model. Social recovery, smart accounts, encrypted backup. The seed-phrase-or-nothing approach loses more users than it protects.
Solve gas before it becomes a support ticket. Sponsored first transactions cost very little and remove one of the most common dead ends.
Support the on-ramp properly. Multiple providers, clear country coverage before the user starts, and honest error messages when a payment fails. Most wallets show a generic failure and lose the user permanently.
Give them a second reason. A single-purpose wallet gets deleted. Staking, swaps, a portfolio view, notifications when something happens — anything that produces a reason to open the app when there’s no transaction to make.
What drives the work
- Chains supported, since each adds integration, indexing and support surface.
- Custody model. Self-custody, MPC and smart accounts have very different onboarding flows and very different support loads.
- Fiat coverage, which is the largest driver of both cost and conversion in most markets.
- Regions and languages, since wallet trust is intensely local.
- What you are permitted to offer in each market, which affects both the product and how you can promote it.
- Whether you have a distribution partner or are acquiring entirely cold, which changes the economics more than any other single factor.
The honest recommendation
Fix the funnel before you buy more installs. Instrument every step, especially the on-ramp failures, and find the specific screen where people leave.
Then buy installs from the narrowest, highest-intent channels you can find, and accept a higher cost per install in exchange for a much lower cost per funded wallet.
The number that goes in the board deck should be funded and retained. If the install number is the headline, someone is being managed rather than informed.
Common questions
Why do most wallet installs never get funded?
Because installing is free and funding requires a decision. The gap is usually a mix of friction — an on-ramp that rejects the user's country or card, a seed phrase screen that frightens people, gas requirements nobody explained — and absence of reason, where the user downloaded the wallet for one specific purpose and has no second one. Fixing the funnel between install and first transaction beats buying more installs almost every time.
What is the right primary metric for a wallet product?
Funded, retained wallets — accounts that received value and were still active 30 days later. Installs, connections and registered accounts are diagnostic. A wallet with two million installs and forty thousand funded accounts has a funnel problem that no additional install spend will solve, and reporting on the larger number hides exactly the thing you need to see.
How do you market a self-custody wallet without scaring people?
Lead with the specific thing the user came to do, not with custody philosophy. People download wallets to buy something, join something or claim something. Explain seed phrase responsibility clearly at the point it matters, offer recovery options that do not compromise the model, and do not turn the first screen into a lecture about sovereignty. The security conversation lands better after the user has a reason to care.
Do wallet airdrops and quests work?
They reliably produce installs and reliably produce poor retention. Quest participants optimise for the reward and leave when it stops. There are cases where the volume is worth it — bootstrapping a network effect, seeding an initial user base before a launch — but the case should be made explicitly, with the retention cost priced in rather than hidden behind an install number.
Which channels work for wallet acquisition?
Distribution partnerships with the apps and protocols your users already use, search on high-intent terms, creator content in the regions you are targeting, and app store optimisation, which is unglamorous and consistently underrated. Broad paid install campaigns produce the worst quality of any channel in this category, because they select for people responding to an ad rather than a need.
