Across the investment materials NarrEx has been tested on, five narrative-model misalignment patterns appear with regularity. Each represents a distinct structural gap between what a company claims in its investment materials and what its financial model actually supports.
These patterns are not unique to any single sector or stage. They recur because they reflect common dynamics in how early-stage businesses think about and present their trajectory. Understanding them helps investors know where to look — and what questions to ask.
This essay goes deeper than labels. For each pattern, we walk through what typically appears in the deck, what is missing or contradictory in the spreadsheet, the economic mechanism that makes the gap costly, the questions a disciplined IC should ask and the benign explanations that sometimes resolve the tension without anyone pretending the materials already match.
If you are skimming, read the pull-quote summaries. If you are diligencing this week, read your pattern twice and bring one page of written questions to the partner meeting. The goal is not to win an argument with management; it is to align underwriting with the business the model actually prices.
1. The Ghost Upsell
The investment materials lead with expansion revenue. Land-and-expand is described as a core growth driver. A specific percentage of net new ARR is attributed to existing customer expansion.
The model contains no expansion logic. ARR is built entirely from new customer acquisition at flat ARPU. There is no cohort-level expansion assumption, no module adoption schedule and no commercial capacity allocated to the upsell motion.
The gap matters because expansion and new logo revenue have fundamentally different economics. When the model does not contain the expansion driver the narrative describes, the projected unit economics are built on a different business than the one being presented.
In diligence, Ghost Upsells often hide in plain sight because expansion language feels “standard” for SaaS. Everyone expects land-and-expand vocabulary. The error is assuming vocabulary implies model structure. A serious pass opens the ARR bridge and asks: where does net retention or module attach rate enter the forecast as a driver, not as a footnote?
Teams sometimes protest that the model is “conservative on purpose.” That can be legitimate. The disciplined response is to separate two questions: (1) is expansion intentionally excluded from the base case, and (2) is the base case still being compared to a valuation narrative that assumes expansion is central? If the answer to both is yes, you do not have a conservative model; you have a mismatched underwriting stack.
Ask for a simple bridge from the percentage of growth attributed to expansion in the deck to a line or cell in the model. If the bridge requires verbal gymnastics, treat expansion as unpriced option value until the next model revision. Option value can be valuable—but it should carry a different discount rate than contracted recurring revenue.
Founders rarely set out to mislead. More often, the commercial team believes expansion will arrive, finance has not yet encoded the motion and the deck merges the two timelines into a single confident story. The fix is operational: assign an owner to reconcile narrative and forecast before external materials ship. Investors can accelerate that fix by refusing to reward smooth decks that are not spreadsheet-backed.
Longer-term, Ghost Upsells are a leading indicator of forecasting maturity. Companies that repeatedly pitch motions the model cannot produce tend to struggle with quota planning, customer success staffing, and module pricing discipline. The diligence artefact is not only the contradiction; it is the organisational gap between go-to-market storytelling and finance integration.
2. The Scaling Paradox
The investment materials describe an efficient, organic growth motion. Sales is described as driven by inbound, referral and partner channels. CAC is stated to be declining.
The model shows that revenue growth is entirely dependent on increasing sales and marketing spend. The implied CAC is rising, not falling. The headcount plan shows accelerating sales hiring. There is no organic or partner-channel assumption embedded in the model.
The narrative claims efficient scaling. The model shows spend-dependent scaling. These are contradictory business models, not different descriptions of the same one.
The Scaling Paradox is emotionally appealing because every investor wants to believe in efficient growth. Inbound-led, partner-led, and community-led stories signal quality. Spend-led stories signal dilution. The model, however, cannot lie about cash: if pipeline conversion requires more reps and more marketing dollars each year, CAC dynamics are unlikely to improve on their own.
A practical test is to trace channel attribution from language to assumptions. If “50% of pipeline is inbound,” the demand-gen sheet should not show 90% of qualified pipeline sourced from outbound SDR activity unless definitions are explained. If “CAC is falling,” the model should show improving conversion or decreasing touch intensity, not only decreasing reported CAC because the denominator changed.
ICs should be willing to hold two truths at once: the company may be building toward efficient channels while still currently dependent on expensive ones. The mistake is letting the deck speak only in the future tense while the model prices the present tense without comment.
When the paradox is real and unresolved, payback periods and burn multiples become fragile. Small changes in conversion or hiring timing swing outcomes because the forecast is levered to spend, not to organic leverage. Stress tests should explicitly break the paradox: what happens if inbound contribution stalls for three quarters while hiring continues on plan?
Resolution paths are straightforward when teams embrace them: either embed partner and inbound coefficients with explicit ramp schedules, or rephrase the narrative until the model catches up. What does not work is leaving heroic channel mix in the prose and brute-force S&M in the spreadsheet.
3. The Lagging Hire
The investment materials project significant revenue acceleration in the next 12 months. The narrative attributes this to expanding into new geographies, launching new product lines, or moving upmarket.
The headcount plan shows that the hiring required to support this expansion does not begin until six to twelve months after the revenue acceleration is projected to start. Sales capacity, customer success, and implementation resources are not in place when the revenue is supposed to arrive.
Revenue cannot scale ahead of the capacity to generate and support it. When the model shows revenue acceleration before the corresponding hiring, the projections are implicitly assuming output that the team does not yet have the capacity to produce.
Lagging Hire patterns are easy to miss because headcount tables look boring next to revenue charts. Yet headcount is where abstract growth becomes operational reality: ramp time, manager span, onboarding classes, territory splits and implementation queues all live there. If revenue jumps a quarter before AE hiring inflects, someone is assuming productivity miracles.
Do not limit the review to sales. Customer success, solutions engineering, implementation and support often bind enterprise and mid-market motions before sales capacity does. A land-and-expand story with flat CS hiring is a different kind of lag—and often the one that breaks NRR first.
Ask for a quarterly map: roles hired, roles fully ramped, and revenue booked, on the same timeline. Misalignment between the three is normal during transitions; persistent separation is not. If management cannot produce the map, they may still be running the business well, but they are not yet running it in a way you can diligence.
Benign explanations exist: planned contractor surge, outsourced implementation partner, or temporary utilisation stretch. Each should leave fingerprints in opex, vendor lines, or backlog metrics. If the only “explanation” is confidence, classify the acceleration as contingent until evidence attaches.
Post-investment, Lagging Hire gaps are leading indicators of forecast credibility. Companies that repeatedly miss hiring plans while maintaining revenue guidance often have forecasting processes that optimise for board optics rather than capacity physics. That is fixable, but only if investors name the pattern early.
4. The Enterprise Illusion
The investment materials describe an upmarket transition. The narrative positions the company as moving from SMB to mid-market or enterprise. Higher ACVs are described as the primary driver of ARR growth.
The model shows flat or slowly growing ARPU. The customer count is growing at a rate that implies continued SMB acquisition. There is no enterprise-specific product, pricing, or sales motion in the model assumptions.
The Enterprise Illusion is particularly common in Series A materials because the upmarket aspiration is real — the team genuinely intends to move upmarket. But intention is not the same as a modelled assumption. Until the enterprise motion is reflected in the model, the ARR projection is implicitly being generated by the SMB business the investment materials are claiming to move away from.
Enterprise transitions change sales cycle length, discount authority, legal review, security posture, professional services attachment, and renewal negotiation dynamics. None of that is free. If ARPU in the model inches upward at 3% while the deck describes a wholesale ICP shift, you are not diligencing the same company.
Useful IC questions: show the last twenty deals closed above your enterprise threshold; show average discount, services attach and sales cycle for that slice; show how those parameters differ from the long tail. Then open the revenue build and ask which of those parameters are explicit inputs. Silence in the model is not neutrality; it is an implicit bet that the SMB build still applies.
The Illusion sometimes resolves when teams admit they are in a hybrid phase: still closing SMB volume while experimenting with enterprise. That honest framing allows underwriting to weight two motions explicitly. The dangerous version is rhetorical enterprise positioning with a forecast that still behaves like mid-market velocity.
Sector overlays matter. Security software, vertical SaaS and regulated workflows often require services-heavy enterprise adoption. If services revenue is buried in “other” while the narrative emphasises pure software ACVs, you may be looking at margin and delivery risk the model does not price.
Longer essays could be written on win-rate math alone. Suffice it here: if enterprise deals are lumpy and slow, pipeline coverage ratios calibrated on SMB motion will mislead every quarter until someone rebuilds the forecast engine. Investors who catch the Illusion early save themselves from surprise Q4 misses that were structurally predictable.
5. The Retention Friction
The investment materials highlight strong retention. Net revenue retention is cited. Customer lifetime is described as long. Churn is described as low and declining.
The cohort data in the model tells a different story. Early-period churn is elevated. The retention curve shows a steepening decay in months six to eighteen before stabilising. The NRR figure cited in the investment materials reflects the stabilised cohort performance, not the current acquisition cohort trajectory.
The Retention Friction matters because it affects the CAC payback calculation. If a business is acquiring customers at a rate that implies a certain payback period, but the actual retention curve means those customers churn before payback is reached, the unit economics are structurally negative in ways the headline numbers do not reveal.
Retention storytelling has its own vocabulary tricks: logo churn versus revenue churn, gross retention versus net, cohorts defined on revenue versus seats and “steady-state” cohorts that exclude the most recent vintage. None of these choices is automatically wrong. The failure mode is comparing a flattering definition in the deck to a harsher curve in the model without reconciling the gap.
Ask explicitly which cohort the cited NRR references and whether that cohort matches the customer population driving current new ARR. Fast-growing companies often outrun their own stabilised cohorts: the headline retention metric ages nicely while the newest customers look meaningfully worse.
Early churn spikes can be a product of ICP experimentation, onboarding gaps, or seasonal selling. Again, benign explanations exist—but they belong in the model as revised assumptions, not only in verbal Q&A. If early-life churn is elevated, payback and LTV math should reflect it until proven otherwise.
For IC discussion, one blunt question often clears the air: “If we acquired no new logos for four quarters, what happens to revenue, and does the model’s churn schedule match that answer?” Founders sometimes pause because nobody has wired the narrative back to that stripped-down scenario.
Retention Friction is where diligence meets capital allocation discipline. A company can grow impressively while destroying value per customer if acquisition tempo masks decay. The model is supposed to expose that tension. When it does not, investors must supply the scrutiny the spreadsheet failed to encode.
How the patterns show up in a typical deck order
Most Series A packs follow a familiar choreography: problem, product, traction, go-to-market, financials, plan, ask. The five misalignments rarely announce themselves on the title slide. They accumulate as you move from story to arithmetic. Experienced readers sometimes work backwards: start from the revenue build and only then return to the GTM chapter, because the build exposes which motions are actually priced.
Early in the deck, you will often encounter the Enterprise Illusion in positioning language—ICP diagrams, named logos, and aspirational ACV commentary—before you have seen any ARPU trajectory. Note the claims mentally, then hold them open until the model tab loads. Ghost Upsell language tends to appear in the same chapters: expansion as strategic pillar, land-and-expand diagrams, net retention called out as a proof point.
The Scaling Paradox usually lives in the tension between marketing copy and S&M trajectory. Inbound and partner slides can sit adjacent to a sales hiring table that tells a different story. Lagging Hire shows up most clearly when revenue inflection is front-loaded in charts but people plans ramp late; sometimes the disconnect is only visible when you overlay quarters on a single page.
Retention Friction is often buried in an appendix cohort chart or a single NRR callout on a summary metric slide. The headline number looks excellent; the shape of the curve requires squinting. If your diligence rhythm skips appendices until the last night, you will miss Retention Friction repeatedly. Promote cohort review earlier in the sequence than feels natural.
None of this implies that founders order slides to obscure gaps. The ordering reflects sales logic: lead with vision, end with numbers. Diligence logic is inverted: let numbers discipline vision before vision seduces the room. A practical compromise is to run two passes—one narrative-forward for context, one model-forward for falsification.
Junior team members can contribute enormously here by time-stamping contradictions as they appear (“slide 14 says X, model assumes Y”). The senior mistake is smoothing those notes away to preserve meeting momentum. Momentum is the enemy of structural clarity.
When two or three patterns stack
Real materials often contain more than one misalignment. Ghost Upsell plus Enterprise Illusion is a frequent pairing: the company describes enterprise expansion motion while the model still prices SMB new logos without expansion drivers. Scaling Paradox plus Lagging Hire appears when spend accelerates but capacity arrives late—effectively a forecast that assumes both efficiency gains and miraculous productivity until hiring catches up.
Stacked patterns are not a verdict on integrity. They are a verdict on forecasting maturity and organisational load. A team juggling multiple motions may simply not have had time to rebuild the model. The investor question becomes whether the round is priced for the messy transitional state or for the cleaned-up future state.
When patterns stack, resist the temptation to negotiate each one in isolation. Ask for a single integrated bridge: hiring, channel mix, ARPU, retention and S&M spend on one timeline. If that integrated bridge cannot be produced, your risk is not only each gap individually but the unknown correlation between them—how a miss on hiring interacts with a miss on retention when both are assumed to improve.
Some ICs use a simple rule: more than two unresolved structural gaps triggers a model rebuild requirement before final approval, or a valuation adjustment that prices manual diligence labour explicitly. The rule is arbitrary but useful; without a trigger, committees defer hard synthesis until after close.
Partner-meeting questions you can borrow verbatim
Questions do not need to be clever. They need to be precise. For Ghost Upsell: “Show me the cell where expansion ARR enters the forecast, or confirm it is intentionally zero in the base case while the deck attributes a specific share of growth to expansion.” For Scaling Paradox: “If inbound is half of pipeline, show the inbound coefficient in the demand model and the implied S&M required if inbound underperforms for two quarters.”
For Lagging Hire: “Overlay this quarter’s revenue target with fully ramped headcount by role; tell me who is selling and supporting the revenue that is supposed to close in Q2.” For Enterprise Illusion: “Name the last ten deals above $X; show cycle time, discount and services attach; then show where those parameters live as inputs.” For Retention Friction: “Which cohort defines the NRR on slide eight and show the same definition applied to customers acquired in the last four quarters.”
The magic is not the wording. It is the expectation of a bounded answer. Open-ended “tell us about retention” invites narrative repair. Bounded requests invite artefacts. Artefacts are what diligence can audit six months later when memories diverge.
If management pushes back on precision, treat that as information. Serious operators welcome bounded questions because they reduce wasted cycles. Teams that resist often do so because the internal reconciliation work is unfinished—which is exactly what you wanted to learn before wiring capital.
Finally, leave room for human judgement. Patterns are heuristics, not laws. A brilliant team with an ugly model may still be the right bet at the right price. The point of naming misalignments is to stop confusing eloquence with evidence, not to replace judgement with checklists.
What good remediation looks like
The best responses to a surfaced misalignment sound boring in a good way: they produce spreadsheets, not speeches. For Ghost Upsell remediation, you want to see an expansion line or module schedule with owners, timing and linkage to quota capacity. For Scaling Paradox remediation, you want channel coefficients or an explicit statement that the company is spend-led with a revised story to match.
Lagging Hire remediation should rearrange either revenue or hiring so the timeline is physically plausible, plus a note on ramp assumptions. Enterprise Illusion remediation shows ARPU, cycle time, and services attach as explicit inputs, even if early estimates are wide. Retention Friction remediation aligns cohort definitions across deck and model and, ideally, shows payback under the pessimistic early curve, not only under stabilised cohorts.
Speed matters. Teams that can turn a first-pass contradiction into an updated model within days are demonstrating forecasting culture. Teams that promise reconciliation “after the round” are asking you to price uncertainty twice—once in valuation and again in post-close surprises.
Investors sometimes worry that pressing for model fixes will poison relationships. The opposite is often true when tone stays technical. Founders are exhausted by vague scepticism; they respect specific requests that reduce ambiguity. The five patterns give you a vocabulary that is about structure, not character.
Document what you received. If remediation artefacts are clean enough, your future self will thank you when the company raises again and new investors ask what you knew and when. If remediation never arrives, your documentation becomes the basis for honest reserve and support decisions.
Why Series A is the natural habitat for these patterns
Later-stage companies usually employ full finance functions with model owners and audit-style forecast discipline. Seed-stage companies are often too early for detailed revenue builds; investors underwrite people and market more than spreadsheets. Series A sits in the uncomfortable middle: enough revenue history to pretend the model is precise, not yet enough organisational muscle to guarantee it is.
Series A is also where narrative professionalisation spikes. Decks hire designers; language converges on category templates; metrics get polished. Models lag because rebuilding them is slow and because no one gets rewarded internally for spreadsheet aesthetics. The five patterns are partly a story about that temporal lag.
None of this implies the patterns disappear entirely at B or beyond. They shrink when forecasting maturity improves and when repeated investor scrutiny raises the cost of unmaintained contradictions. Until then, the same ghosts reappear with nicer fonts.
If you work at growth stage, you can still use this frame as a rapid triage tool when diligencing acquisitions or when a portfolio company suddenly refreshes its story ahead of an inside round. The vocabulary scales; only the tolerance for slack changes.
A longer afternoon: pairing this essay with the model tabs
Print the five section headings on a single page. Open the model in a second window. For each heading, spend twenty minutes hunting disconfirming evidence before you hunt confirming narrative. Write one bullet per pattern: not observed / possible / likely / clearly present. If you finish in less than an hour, you were either looking at an unusually clean pack or skimming.
Bring those bullets to the partner meeting unchanged. Let the group argue about likelihood, not about memory. The goal of the exercise is to prevent the meeting from becoming a retelling of the deck. Deck retellings are how misalignments survive.
When you are done, file the bullets in the deal folder. Six months later, compare them to board materials. That loop—predicted structural risk versus observed operating reality—is how investors calibrate their own diligence quality over time, independent of any single outcome.
Patterns are training wheels. The end state is instinctive scepticism about unpriced claims, without cynicism about operators. The best investors we know sound almost dull when they describe how they read materials: they check bridges, definitions and timelines the way pilots check instruments. This essay is simply a longer way of saying: fly with instruments, not with slogans.
Where to click: a rough map of the workbook
Models vary, but revenue builds usually announce themselves. Look for ARR or MRR bridges, new versus expansion versus churn lines and net retention inputs. If expansion is narratively central, absence of a line is meaningful even when logos look enterprise-grade. If net retention is cited without a driver table, ask where the driver lives.
Sales and marketing tabs are where Scaling Paradox hides. Scan S&M as a percentage of revenue, headcount by quarter and any pipeline-to-revenue conversion assumptions. Inbound-led stories should not require endless S&M step-ups unless the model explicitly encodes a temporary investment phase with an end date.
People plans are rarely glamorous sheets, yet they adjudicate Lagging Hire claims faster than most narrative paragraphs. Compare ramp dates for AEs, SEs, CS and implementation against revenue inflection timing. If the model uses productivity per rep, check whether productivity is assumed to jump precisely when hiring is thin.
Customer count and ARPU splits address Enterprise Illusion quickly when they exist. If only logo count grows while ARPU flatlines, the enterprise story is mostly rhetorical until proven otherwise. If ARPU grows, verify whether growth is price, mix, discount, or one-off upsells—each has different persistence.
Cohort tabs and churn schedules are Retention Friction terrain. Read vintage curves, not only summary NRR. Pay attention to definitions of churn events, revenue versus seat churn, and whether early months are excluded. Small definition shifts can flatter a headline while leaving economics harsh.
Cash and working capital sheets sometimes contradict efficiency narratives indirectly. If working capital stretches while the story emphasises software-like cash conversion, ask why. Not every contradiction is nefarious; some are seasonal or growth-related. The point is to surface the tension early.
Scenario tabs, if present, deserve a skim even when you underwrite base case only. Wide scenario spreads with narrow narrative confidence often indicate internal disagreement that did not make the deck. Ask which scenario management actually believes and why the base case landed where it did.
Finally, version control matters. If you receive multiple model versions during a process, diff them. Silent changes between v3 and v4 have altered more deals than dramatic partner debates. A simple change log from the company saves everyone time.
If you are new to financial models, pair with someone who is not. Narrative readers catch language slips; spreadsheet readers catch unit errors. The five misalignments sit at the intersection; you want both skills in the room before conviction sets.
Treat this map as a starting orientation, not an exhaustive audit programme. The deeper point is that models are not neutral containers. They encode a theory of how value is created. Your job is to compare that theory to the theory implied by the prose. When the two diverge, you have found work worth doing—whether or not you end up investing.
Some teams will never produce perfect alignment at Series A, and that is acceptable when price and structure reflect the gap. What is not acceptable is pretending alignment exists because the deck is handsome. Handsome decks are cheap. Honest bridges between language and arithmetic are expensive—and that is why they are worth paying attention to.
We will close where we began: these five patterns are frequent, not universal. They are descriptive, not moral. They are a lens for reading, not a substitute for judgement. Used well, they help you spend less time on theatre and more time on the specific work that turns conviction into fair pricing—and fair pricing into durable partnerships with founders.
A note on tone: diligence without contempt
It is possible to read this essay and walk into a founder meeting sounding like a prosecutor. That is a failure mode. The patterns describe structural tension between two artefacts—deck and model—not moral failure. Teams under pressure produce tension; your job is to reduce ambiguity, not to score rhetorical points.
The most effective questions are delivered as shared problem-solving: “Help us reconcile these two views so we can underwrite accurately.” That framing invites collaboration. By contrast, “You lied about expansion” invites defensiveness and hides information even when the underlying gap is real.
Founders deserve clarity about what would resolve concern. If the answer is a model revision, say so. If the answer is a price change, say so. If the answer is simply time and proof, say that too. Opaque scepticism wastes everyone’s calendar.
Investors deserve clarity about what they are pricing. If you choose to invest despite a known Ghost Upsell because you believe in the team’s ability to operationalise expansion post-close, own that as a thesis bet. Do not pretend the model already contains what it lacks. Intellectual honesty compounds trust across rounds.
Finally, remember that founders read your behaviour as a signal of post-close partnership. Harsh diligence paired with supportive, precise follow-through builds reputations. Harsh diligence paired with vague noes does not. The five patterns are a tool for specificity—and specificity is a form of respect.
If you are a founder reading this from the other side of the table: the patterns are not an indictment. They are a description of what happens when commercial storytelling moves faster than finance integration. The fastest way through a sharp diligence process is often to pre-reconcile deck and model before materials ship, then volunteer the tension points yourself. Surprises discovered by investors feel worse than tensions disclosed upfront.
If you are an investor racing a deadline: pick two patterns most material to the sector and spend your deep time there rather than spreading thin. Depth beats breadth when hours are scarce. Return to the others on a second pass if the process continues.
That is the full tour: definitions, mechanics, IC prompts, stacking behaviour, remediation, workbook orientation, tone, and practical pacing. Keep the printout; reuse it until the questions become reflexive. The patterns will still appear in materials long after this essay is outdated—because the underlying human incentives persist even as slide templates change.
Use it as a long companion piece, not a checklist to rush. The best diligence weeks mix slow reading with fast questions: slow enough to notice contradictions and fast enough to give founders time to respond well. If this essay helps you strike that balance even once, it has done its job.
Return to it when a deal feels “too easy”—that is often coherence doing its job on your intuition. Easy reads deserve the same confirmation pass as hard ones; the cost is low and the upside is a cleaner conscience at allocation.
Each of these five misalignments shares a common structure: the narrative describes a business that is more efficient, more scalable, or better positioned than the model actually supports.
What these patterns have in common
In each case, the gap is not necessarily the result of deliberate misrepresentation. It reflects the natural tendency for commercial narrative to outrun modelled assumptions.
They also share a social dynamic: by the time materials reach investors, many internal stakeholders have already repeated the same phrases in meetings, emails, and slide reviews. Repetition creates felt truth. The model, unsentimental as it is, becomes the last place where the story is still forced to be explicit.
That is why the most productive IC conversations are rarely adversarial. They treat the spreadsheet as a translation layer: if the translation is missing, the work is to complete it, not assign blame. Founders who experience diligence as collaborative translation tend to return with stronger materials; founders who experience it as character judgement tend to optimise for opacity.
NarrEx detects these patterns by mapping each narrative claim to its corresponding model evidence and flagging where the claim is not supported. The goal is not to disqualify businesses that exhibit these patterns — many strong businesses have narrative-model gaps that are easily explained and quickly resolved. The goal is to surface the gaps early enough that the right questions can be asked before conviction hardens into consensus.
If you take one habit away from this piece, make it a small one: before every partner meeting, write down the single claim you would most hate to discover was unsupported six months post-close. Then open the model and search for that claim the way an auditor would—not the way a storyteller would. The five patterns above are where that search most often lands in Series A materials.
These five patterns represent the most frequently detected misalignments in NarrEx’s current claim taxonomy. The taxonomy will expand as the platform moves from revenue-focused validation into operational and market narrative assessment.