A lever board is only as good as its causal model. This is the model, lever by lever: how the effect was measured, the parameters with their ranges, and the published research behind the rule. Every number is invented for this page; the shape of the evidence is what a real one would need.
How it was measured: the finance plan for the year, built from last year's spend and the sales system.
| The plan | $1M a year, all-in |
| Range of the lever | $10 000 to $10M a year |
| How it splits at the plan | $815 000 on people, $185 000 on programmes |
| A person, fully loaded | $200 000 a year in a sales team, $215 000 for an integration engineer, tools and home-city travel included |
| New hires | reach full speed after 6 months |
| Who the budget hires | people come whole or by halves: the budget hires them half a person at a time, in this order, until people take about their plan share of it: institutions, curators, partners, integration, banks, partners, integration, chains, fixed rate; below $100 000 a year it cannot pay half a person, and nobody works the meetings |
Studies: Personal Selling Elasticities: A Meta-Analysis, Dragonfly 2024/2025 Crypto Compensation Report, 2026 SaaS and AI Metrics Benchmarks.
In production this lever reads from the budget sheet and the HR system.
How it was measured: requests through the docs, the app and the team's public channels, tagged at first contact.
| Requests a month | 6 partners, 3 institutions, 2 banks, 4 curators, 1 chains, 0.5 fixed rate |
| By intent | to launch an earn or borrow product: 7.5 a month; to curate: 4 a month; to embed the protocol in their stack: 2 a month; to lend or borrow directly: 3 a month |
| Who answers them | the team of the motion each request belongs to; a full team works the best of what arrives and the rest wait and go cold: at the plan, about 24 qualified requests a year wait |
| The what-if | from 0.5 to 1.5 times the plan, costs nothing on the board |
Studies: Consumer Heterogeneity and Paid Search Effectiveness: A Large-Scale Field Experiment, Close Enough? A Large-Scale Exploration of Non-Experimental Approaches to Advertising Measurement, The Unfavorable Economics of Measuring the Returns to Advertising.
In production this lever reads from the sales system and the request form.
How it was measured: meetings logged against each trip in the sales system, keeping only counterparties not met before.
| Spend in the plan | $54 000 a year, about 9 person-trips at $6 000 each |
| First meetings | 12 (range 6 to 25) for each person-trip, 108 a year at the plan |
| How often a meeting qualifies | 1 times the average, so about 30 percent of these meetings become a qualified opportunity |
| Response to more spend | each extra person-trip brings less than the one before: twice the spend brings about 1.80 times the meetings |
| Who it meets | 30 percent partners, 20 percent institutions, 10 percent banks, 25 percent curators, 10 percent chains, 5 percent fixed rate |
Studies: A Field Experiment on Search Costs and the Formation of Scientific Collaborations, Personal Selling Elasticities: A Meta-Analysis.
In production this lever reads from the sales system and the travel budget.
How it was measured: badge scans matched to meetings booked within two weeks.
| Spend in the plan | $15 000 a year, about 5 booths at $3 000 each |
| First meetings | 5 (range 2 to 10) for each booth, 25 a year at the plan |
| How often a meeting qualifies | 0.6 times the average, so about 18 percent of these meetings become a qualified opportunity |
| Response to more spend | each extra booth brings less than the one before: twice the spend brings about 1.68 times the meetings |
| Who it meets | 35 percent partners, 20 percent institutions, 5 percent banks, 25 percent curators, 15 percent chains |
Studies: Close Enough? A Large-Scale Exploration of Non-Experimental Approaches to Advertising Measurement, The Unfavorable Economics of Measuring the Returns to Advertising.
In production this lever reads from the event platform and the sales system.
How it was measured: contacts from each sponsorship matched to the sales system for 60 days, and the share of first meetings with people who already knew the name.
| Spend in the plan | $50 000 a year, about 1 sponsorship at $50 000 each |
| First meetings | 20 (range 8 to 40) for each sponsorship, 20 a year at the plan |
| How often a meeting qualifies | 0.8 times the average, so about 24 percent of these meetings become a qualified opportunity |
| Response to more spend | each extra sponsorship brings less than the one before: twice the spend brings about 1.62 times the meetings |
| Who it meets | 35 percent partners, 20 percent institutions, 10 percent banks, 25 percent curators, 10 percent chains |
| Builds the brand | one more sponsorship a year makes the brand about 5.0 percent stronger by month twelve (range 2.0 to 10.7), building through the year |
Studies: The 5 Principles of Growth in B2B Marketing, Close Enough? A Large-Scale Exploration of Non-Experimental Approaches to Advertising Measurement.
In production this lever reads from the event platform, the sales system and the follower counts.
How it was measured: guest lists matched to the sales system 30 days after each event.
| Spend in the plan | $15 000 a year, about 1 side event at $15 000 each |
| First meetings | 12 (range 6 to 18) for each side event, 12 a year at the plan |
| How often a meeting qualifies | 0.8 times the average, so about 24 percent of these meetings become a qualified opportunity |
| Response to more spend | each extra side event brings less than the one before: twice the spend brings about 1.74 times the meetings |
| Who it meets | 35 percent partners, 20 percent institutions, 10 percent banks, 20 percent curators, 10 percent chains, 5 percent fixed rate |
Studies: A Field Experiment on Search Costs and the Formation of Scientific Collaborations, Close Enough? A Large-Scale Exploration of Non-Experimental Approaches to Advertising Measurement.
In production this lever reads from the event platform and the sales system.
How it was measured: an invitation holdout: half of each target list invited at random, the other half not.
| Spend in the plan | $25 000 a year, about 5 dinners at $5 000 each |
| First meetings | 7 (range 4 to 10) for each dinner, 35 a year at the plan |
| How often a meeting qualifies | 1.4 times the average, so about 42 percent of these meetings become a qualified opportunity |
| Response to more spend | each extra dinner brings less than the one before: twice the spend brings about 1.87 times the meetings |
| Who it meets | 35 percent partners, 25 percent institutions, 15 percent banks, 20 percent curators, 5 percent chains |
Studies: A Field Experiment on Search Costs and the Formation of Scientific Collaborations, Personal Selling Elasticities: A Meta-Analysis.
In production this lever reads from the guest lists and the sales system.
How it was measured: meetings held on three city trips in 2026; there is nothing to compare them with.
| Spend in the plan | $10 000 a year, about 2 roadshows at $5 000 each |
| First meetings | 8 (range 5 to 12) for each roadshow, 16 a year at the plan |
| How often a meeting qualifies | 1.2 times the average, so about 36 percent of these meetings become a qualified opportunity |
| Response to more spend | each extra roadshow brings less than the one before: twice the spend brings about 1.74 times the meetings |
| Who it meets | 25 percent partners, 35 percent institutions, 30 percent banks, 10 percent curators |
Studies: Personal Selling Elasticities: A Meta-Analysis, Drivers of Sales Performance: A Contemporary Meta-Analysis. Have Salespeople Become Knowledge Brokers?, AE Models, Motions & Metrics: 2026 Research Report.
In production this lever reads from the travel budget and the sales system.
How it was measured: the protocol's rank among lending protocols by followers of each official X account, read on TwitterScore; what a rank does to the other levers comes from first meetings and win rates by how the counterparty knew the name.
| This protocol in the plan | #8 of 13, between Euler and Spark: a reach of about 69 000 followers |
| What a place is worth | strength is reach against the plan's, to the power 0.15; between two protocols, reach is the geometric mean of the two. #1 makes it 1.45 times today's, #4 1.13, #13 0.91 |
| What strength moves | first meetings from every source with the strength itself, the share that qualifies and the win rate with its square root: at #1, first meetings +45 percent, qualify and sign +20 percent each |
| Sponsorships | one more sponsorship a year, from the plan's 1, makes the brand about 5.0 percent stronger by month twelve (range 2.0 to 10.7), building through the year |
| # | Lending protocol | X followers |
|---|---|---|
| 1 | Aave | 700 914 |
| 2 | Compound | 255 495 |
| 3 | Venus | 223 084 |
| 4 | Radiant | 105 000 |
| 5 | JustLend | 78 754 |
| 6 | Benqi | 72 473 |
| 7 | Euler | 70 806 |
| 8 | Spark | 66 607 |
| 9 | Maple | 62 411 |
| 10 | Moonwell | 54 765 |
| 11 | Silo | 45 500 |
| 12 | Fluid | 40 855 |
Read on TwitterScore on 20 October 2026. Followers include bots and old token audiences: a measure of reach, not a survey. JustLend's count is the JUST DAO account, which covers its other products too.
Studies: The 5 Principles of Growth in B2B Marketing, A Field Experiment on Search Costs and the Formation of Scientific Collaborations.
In production this lever reads from the follower counts each month, and the sales system.
How it was measured: deposits and loans through each partner's own vault, counted on-chain.
| Team in the plan | 1 person, $200 000 a year |
| Counterparties worth a meeting | about 350: exchanges, wallets, fintechs and neobanks that put earn and borrow inside their app; a second meeting with the same one is not a first meeting |
| Outbound | 3 (range 2 to 4.5) first meetings a month for each person |
| Qualified | 30 (range 20 to 45) percent of first meetings, the same for every motion |
| Capacity | 2.2 qualified opportunities a month for each person; the rest wait and go cold |
| Win rate | 25 (range 15 to 35) percent of worked opportunities sign |
| Time to live | 8 (range 4 to 14) months from qualified to live, then the launch waits for an integration engineer; with discovery before it, about 12 months from the first meeting, the time the Partner time to launch what-if moves |
| Each win, once ramped | $60M of deposits (range $10M to $600M) and $10M of borrowing (range $1M to $800M); an integration: the partner's users deposit through a vault the partner picks; fewer than one in five partners also lets them borrow against their crypto, and those bring most of the borrowing |
| Ramp | 6 months to its full size |
| Counted as caused by the team | 80 (range 50 to 95) percent of the volume |
| At the plan | 139 first meetings, 45 qualified, 6.6 signed in the year |
Studies: Why DeFi Lending? Evidence from Aave V2, DeFi Protocols for Loanable Funds: Interest Rates, Liquidity and Market Efficiency, Close Enough? A Large-Scale Exploration of Non-Experimental Approaches to Advertising Measurement.
In production this lever reads from the vault balances on-chain and the sales system.
How it was measured: an outreach holdout: 60 named institutions, picked at random, left out of outreach for a quarter.
| Team in the plan | 0.5 people, $100 000 a year |
| Counterparties worth a meeting | about 800: funds, trading firms, treasuries and stablecoin issuers that lend or borrow directly, in weeks; a second meeting with the same one is not a first meeting |
| Outbound | 4 (range 3 to 6) first meetings a month for each person |
| Qualified | 30 (range 20 to 45) percent of first meetings, the same for every motion |
| Capacity | 4.5 qualified opportunities a month for each person; the rest wait and go cold |
| Win rate | 30 (range 20 to 40) percent of worked opportunities sign |
| Time to live | 4 (range 2 to 9) months from qualified to live |
| Each win, once ramped | $12.5M of deposits (range $2.5M to $75M) and $7.5M of borrowing (range $2.5M to $50M); half are lenders (about 25M each) and half borrowers (about 15M each), a few of them much bigger; balances leave at 3 percent a month |
| Ramp | 1 months to its full size; balances leave at 3 percent a month |
| Counted as caused by the team | 60 (range 45 to 75) percent of the volume |
| At the plan | 100 first meetings, 32 qualified, 7.7 signed in the year |
Studies: Why DeFi Lending? Evidence from Aave V2, Decentralized Lending and Its Users: Insights from Compound, Anatomy of a Run: The Terra Luna Crash.
In production this lever reads from the whitelisted addresses on-chain and the sales system.
How it was measured: design partnerships followed stage by stage in the sales system; too few have gone live to test.
| Team in the plan | 0.5 people, $100 000 a year |
| Counterparties worth a meeting | about 400: banks, asset managers and custodians that build their own lending product on the protocol as design partners; they need privacy and KYC first; a second meeting with the same one is not a first meeting |
| Outbound | 2 (range 1 to 3) first meetings a month for each person |
| Qualified | 30 (range 20 to 45) percent of first meetings, the same for every motion |
| Capacity | 1.5 qualified opportunities a month for each person; the rest wait and go cold |
| Win rate | 25 (range 15 to 40) percent of worked opportunities sign |
| Time to live | 15 (range 9 to 24) months from qualified to live, then the launch waits for an integration engineer |
| Each win, once ramped | $50M of deposits (range $10M to $500M) and $40M of borrowing (range $5M to $400M); a design partnership: a pilot with one product, then more customers and countries, then privacy, KYC and insurance; most of the volume comes after the twelve months |
| Ramp | 9 months to its full size |
| Counted as caused by the team | 75 (range 60 to 90) percent of the volume |
| At the plan | 57 first meetings, 19 qualified, 2.2 signed in the year |
Studies: Why DeFi Lending? Evidence from Aave V2, Drivers of Sales Performance: A Contemporary Meta-Analysis. Have Salespeople Become Knowledge Brokers?, AE Models, Motions & Metrics: 2026 Research Report.
In production this lever reads from the sales system and the whitelisted addresses on-chain.
How it was measured: vault balances by curator and market balances by collateral, on-chain.
| Team in the plan | 0.5 people, $100 000 a year |
| Counterparties worth a meeting | about 300: risk managers who run vaults, and issuers of collateral such as tokenised treasuries, staked ETH or BTC wrappers; a second meeting with the same one is not a first meeting |
| Outbound | 3 (range 2 to 4) first meetings a month for each person |
| Qualified | 30 (range 20 to 45) percent of first meetings, the same for every motion |
| Capacity | 3.5 qualified opportunities a month for each person; the rest wait and go cold |
| Win rate | 35 (range 25 to 50) percent of worked opportunities sign |
| Time to live | 2 (range 1 to 5) months from qualified to live |
| Each win, once ramped | $16M of deposits (range $4M to $60M) and $24M of borrowing (range $3M to $300M); 4 in 10 wins are curators who bring deposits to a new vault (about 40M each, part of it moved over from other vaults on the protocol); 6 in 10 are issuers whose asset becomes collateral, which brings borrowers (about 40M each once it has ramped, far more when holders can loop it) |
| Ramp | 6 months to its full size |
| Counted as caused by the team | 65 (range 40 to 85) percent of the volume |
| At the plan | 96 first meetings, 31 qualified, 7.2 signed in the year |
Studies: DeFi Protocols for Loanable Funds: Interest Rates, Liquidity and Market Efficiency, Phantom Liquidity in Decentralized Lending, Anatomy of a Run: The Terra Luna Crash.
In production this lever reads from the vault and market balances on-chain and the sales system.
How it was measured: two deployments in 2025 and 2026; deposits followed the chain's incentives.
| Team in the plan | 0.5 people, $100 000 a year |
| Counterparties worth a meeting | about 60: blockchains that want lending on their network and pay incentives to get it; a second meeting with the same one is not a first meeting |
| Outbound | 2 (range 1 to 3) first meetings a month for each person |
| Qualified | 30 (range 20 to 45) percent of first meetings, the same for every motion |
| Capacity | 2 qualified opportunities a month for each person; the rest wait and go cold |
| Win rate | 30 (range 20 to 40) percent of worked opportunities sign |
| Time to live | 4 (range 3 to 6) months from qualified to live, then the launch waits for an integration engineer |
| Each win, once ramped | $100M of deposits (range $20M to $500M) and $35M of borrowing (range $5M to $200M); a deployment on a new chain, with the chain's incentives for six months; once they end, about a quarter of what is left leaves each month |
| Ramp | 3 months to its full size; after 6 months 75 percent stays (range 55 to 88) |
| Counted as caused by the team | 50 (range 30 to 70) percent of the volume |
| At the plan | 31 first meetings, 9 qualified, 2.7 signed in the year |
Studies: STIP Retroactive Analysis: Spot DEX TVL, Airdrops: Giving Money Away Is Harder Than It Seems.
In production this lever reads from the deployment balances on-chain and the chain agreements.
How it was measured: offers quoted and taken on the fixed-rate markets, matched on-chain to the counterparties the desk brought.
| Team in the plan | 0 people, $0 a year |
| Counterparties worth a meeting | about 150: lenders and market makers who quote fixed-term offers, and treasuries, funds and companies that need a rate fixed for the term; a second meeting with the same one is not a first meeting |
| Outbound | 4 (range 2 to 6) first meetings a month for each person |
| Qualified | 30 (range 20 to 45) percent of first meetings, the same for every motion |
| Capacity | 3 qualified opportunities a month for each person; the rest wait and go cold |
| Win rate | 30 (range 20 to 45) percent of worked opportunities sign |
| Time to live | 2 (range 1 to 4) months from qualified to live |
| Each win, once ramped | $4M of deposits (range $1M to $25M) and $12M of borrowing (range $3M to $60M); half are makers who quote 30 to 180 day offers, half takers who borrow at a rate fixed for the term; a taker lends deposits that would otherwise wait, at a premium over the variable rate; terms roll off at about 2 percent a month |
| Ramp | 2 months to its full size; balances leave at 2 percent a month |
| Counted as caused by the team | 70 (range 50 to 85) percent of the volume |
| At the plan | 12 first meetings, 4 qualified, 0.0 signed in the year |
Studies: DeFi Protocols for Loanable Funds: Interest Rates, Liquidity and Market Efficiency, Why DeFi Lending? Evidence from Aave V2.
In production this lever reads from the offers and the loans on the fixed-rate markets on-chain, and the sales system.
How it was measured: the date of every stage of each partner launch in the sales system, from the first meeting to live.
| Today | about 12 months from the first meeting to live for a partner: 8 months from qualified to live (range 4 to 14), discovery before it, and the wait for an engineer |
| The goal | 6 months |
| The what-if | from 6 to 18 months; from month one it moves every partner deal's path to live in proportion, the deals already on it included. It costs nothing on the board: the question is what it would take |
| What holds it back | a launch that arrives sooner still waits for an integration engineer: shorter launches pay only if the engineers keep up |
Studies: Drivers of Sales Performance: A Contemporary Meta-Analysis. Have Salespeople Become Knowledge Brokers?, Personal Selling Elasticities: A Meta-Analysis.
In production this lever reads from the stage dates in the sales system.
How it was measured: launches shipped in 2025 and 2026 against engineer time.
| In the plan | 1 person, $215 000 a year |
| Launches | 1 (range 0.6 to 1.4) launches a month for each engineer |
| A launch that waits | loses 20 percent of its chance each month it waits |
| At the plan | 11.5 launches ready in the year against room for 12.0 |
Studies: Drivers of Sales Performance: A Contemporary Meta-Analysis. Have Salespeople Become Knowledge Brokers?, AE Models, Motions & Metrics: 2026 Research Report.
In production this lever reads from the sales system and the engineering tracker.
How it was measured: co-marketing given by coin flip to half of the 2026 launches.
| In the plan | $16 000 a year at $8 000 a launch, enough for 2 of the 11 launches |
| A supported launch | ends 10 (range 2 to 25) percent bigger and ramps 25 (range 5 to 45) percent faster |
Studies: Personal Selling Elasticities: A Meta-Analysis, Airdrops: Giving Money Away Is Harder Than It Seems, STIP Retroactive Analysis: Spot DEX TVL.
In production this lever reads from the vault balances on-chain and the marketing budget.
How it was measured: incentives switched off in two markets for four weeks.
| In the plan | $1M a year from the treasury, half to depositors and half to borrowers |
| What a dollar holds | one dollar a year holds about 33 dollars of deposits (range 10 to 67) or 40 of borrowing (range 12 to 67) while it runs |
| Counted as caused by them | 37 (range 12 to 62) percent: the rest would have come anyway |
| After they stop | about 5 percent of the peak is still there a year later (range 1 to 15); inside the twelve months they never stop, so the board counts their share of the volume all year and charges their full cost |
Studies: STIP Retroactive Analysis: Spot DEX TVL, Phantom Liquidity in Decentralized Lending, Decentralized Lending and Its Users: Insights from Compound.
In production this lever reads from the incentive payments and the market balances on-chain.
| Assumption | Value | In production, read from |
|---|---|---|
| The protocol today | $7B of deposits and $4.5B of loans; the board counts only what the team adds in the twelve months | the protocol's own data on-chain |
| Interest borrowers pay | 4.5 (range 3.5 to 7) percent a year on the new loans, blended across markets | the markets on-chain |
| Protocol fee | off in the plan, as at many lending protocols while they grow: depositors keep all the interest, and curators and partners keep their own fees. A what-if from off to 25 percent; at 10 percent the protocol would keep a tenth of the interest | governance |
| A rate fixed for the term | borrowers pay 1.5 points more for it than for a variable rate (range 0.5 to 3.0); a fixed-rate book works only when lenders who quote and borrowers who take both show up | the fixed-rate markets on-chain |
| Deposits and the fee | depositors keep the interest less the fee; 10 percent less yield brings about 8 percent fewer deposits (range 1 to 20) | the markets on-chain |
| Lendable | 90 percent of new deposits can be lent before rates jump; $150M of liquidity already in the markets can serve new borrowers (range $50M to $300M) | the markets on-chain |
| Deposits without a borrower | 30 percent of lendable new deposits find a borrower anyway, because rates fall (range 10 to 50); the rest earns nothing | the markets on-chain |
| Crypto prices | a what-if from 0.5 to 1.5 times today; 10 percent higher prices bring about 7.5 percent more borrowing and 2.9 percent more deposits | price feeds |
| Meeting quality | each source qualifies at its own rate: a cold outbound meeting 0.8 times as often as the average, an inbound request 1.3 times, each programme as its card says | the sales system |
| A full team picks | when more qualified opportunities arrive than a team can work, it works the best of them: twice as many to choose from makes its deals about 1.19 times bigger (range 1.07 to 1.32) | the sales system and the chain |
| What-if levers | inbound requests, the meetings that qualify, the win rate, the partner time to launch and the size of each win each move from 0.5 to 1.5 times the plan. They cost nothing on the board, so the question for each is what it would take to move it | |
| Pipeline in flight | the board assumes the team has run at this budget with the plan mix for two years, so deals already in the pipeline go live in the first months; a move you make today pays only after its own lag | the sales system |
| Horizon | twelve months from 2 November 2026. The interest and the revenue are run rates at month twelve; deals signed but not live yet are shown as in flight, not counted | |
| Payback | the sales budget over what it adds a month at month twelve, against the same levers with no budget at all; with incentives, both costs over what both add. With the fee switch off nothing pays it back, so the board shows what a 10 percent fee would do on the interest the budget adds | the accounts |
| The range | the board runs the model 300 times, each time drawing every uncertain number from its own range, and prints where 8 runs in 10 land |
Payback is held against 16 months, the median for software companies; there is no published payback benchmark for a protocol's sales team.