the question nobody wants to ask
you have two projects. one pays the bills and bores you to tears. the other keeps you up at night, the one you open at 11pm when you should be sleeping. you tell yourself the boring one is the real business and the fun one is a hobby. but what if you have it backwards? what if the fun one is the actual business and the boring one is just a very expensive fear blanket?
this is the question that sits under half the creator economy. and for once, there is actual research on it. not vibes. evidence.
the identity trap
a three-year field study of first-time founders published in organization science found something uncomfortable. founder identity is not static. it shifts. and those shifts predict real strategic decisions about whether to broaden or narrow a venture. the same person who says "this is just a side project" one quarter can be all-in the next. the research separates founder identification from psychological distance ... two distinct dimensions that should never be collapsed into one measure of commitment.
the uncomfortable implication: when you call something a side project, you are not describing reality. you are performing a identity. and the label changes what you do next.
another study on founder role identity found the same mechanism from a different angle. founders filter information through their identity lens. an opportunity-discovery founder sees different signals than a business-operation founder, even when looking at the same data. so when you tell yourself the fun project is not viable, ask this: is that a conclusion from evidence, or is it your identity filter doing the thinking for you?
the fear component you cannot see
here is where it gets personal. research on escalation bias among technology entrepreneurs shows that persistence after negative feedback is often not a rational bet. it is a bias. and the founders most susceptible to it are the ones with high intrinsic motivation. in plain language: the more you love the thing, the harder it is to kill it when the evidence says it is not working.
that is the opposite of the side project trap. that is the dream project trap. but there is another side. the same mechanism runs in reverse. you stay in the boring business because leaving it means facing the identity question. who are you without the safe thing? that is not a strategic decision. that is fear wearing a suit.
the research is clear that identity attachment can inhibit adaptation after negative feedback. one babson study found founders who were psychologically attached to a venture's core value proposition were less likely to pivot when the market pushed back. the attachment itself became the obstacle.
what the evidence actually tells you
so how do you build an agent that can tell the difference between a side project and a real business hiding in plain sight? the research points to behavior, not self-report. nobody is going to say "i am keeping this business because i am scared to pursue the other one." but their behavior will say it for them.
here is what the evidence supports:
- watch response to feedback. do you treat negative signals as data or as attacks? identity attachment shows up as defensive interpretation.
- watch resource flows. the escalation bias research found that setting no mental budget strengthens the bias. if the boring project has no predefined spending limit, that is a red flag you are funding a fear, not a business.
- watch stated preference versus revealed behavior. a pricing validation playbook from one vendor documents a simple rule: when survey claims conflict with observed customer behavior, trust the behavior. the same logic applies to your own decisions. what you say you want and what you actually do will diverge. the doing wins.
- watch the longitudinal pattern. the founder identity research supports observing identity over time rather than taking a one-time diagnostic. a single moment of doubt is not a signal. a repeated pattern of choosing safety over the project that lights you up ... that is a signal.
the truth about thresholds
everyone wants a clean rule. if the fun project makes x dollars for three months, it is a business. if it does not, it is a hobby. the research says this is a fantasy. the metrics being thrown around by vendors ... conversion rates, churn, net revenue retention, LTV/CAC ... are heuristics, not validated predictions. they are starting points, not verdicts.
but the absence of a perfect threshold does not mean the absence of evidence. it means you need to build a system of evidence-triggered review points. the escalation bias research is explicit: budgeted, evidence-triggered checkpoints make persistence falsifiable. you do not need to know the exact number that means "quit." you need a predefined moment where you look at the data you committed to collecting and make a call.
the vendor content on churn interviews offers a practical model. real cancellation and downgrade conversations provide revealed-preference evidence that is more diagnostic than a cancellation label or a hypothetical willingness-to-pay survey. strip away the vendor marketing and you have a workflow: connect observed behavior to structured interviews, aggregate patterns, preserve verbatim evidence. then let the evidence speak.
what lunari is building
this is not theoretical. LUNARI is building an agent that watches your actual behavior patterns across projects ... the hours you log, the momentum you sustain, the money that comes in, the feedback you act on. it is not asking you how you feel. it is watching what you do.
the agent looks for the signature signs the research identified. it watches for identity-protective interpretation, the way you explain away negative signals when they come for the dream project. it tracks escalation bias, how long you pour time into a thing after the evidence turns. it measures reversibility, because the research suggests persistence is healthier when it is budgeted and reviewable. and it watches the divergence between your stated preferences and your revealed behavior, because that gap is where the fear actually lives.
does it have all the answers? no. the research base here is partial. much of it is drawn from adjacent domains ... founder identity studies, escalation bias literature, pricing validation frameworks ... not from a direct longitudinal study of side-project promotion versus business fear. there is no validated decision threshold yet. the signal ranking is not established. this is honest.
but the direction is right. you do not need a perfect diagnostic to know that asking someone "are you keeping this business because you are scared?" will not work. nobody answers that question truthfully, because the person asking is the same person who built the fear. you need an external observer that tracks behavior over time and surfaces patterns the identity filter hides.
the takeaway
you will not think your way out of this one. your identity is filtering the information before it reaches your conscious decision-making. your intrinsic motivation is biasing your persistence judgments. your stated preferences are diverging from your revealed behavior and you cannot see it from the inside.
the research supports one practical move: build observable checkpoints. set budgets before you feel the pain. decide what evidence would change your mind before you have gathered it. watch what you actually do, not what you tell yourself you believe.
the fun project might be a hobby. it might also be the real business, and the boring one might be the most expensive fear you ever funded. the only way to know is to stop asking how you feel and start watching what you do.
that is what the agent is for. that is what the evidence supports. and that is the question you have been avoiding.