Investment Thesis Sample: The 40-Word Argument Every SaaS Round Tests

Investment Thesis Sample: The 40-Word Argument Every SaaS Round Tests - hero image

Most founders search­ing for an invest­ment the­sis sam­ple assume they are look­ing for a fill-in-the-blank doc­u­ment. They are not. They are look­ing for the one para­graph that decides whether your Series A clos­es in six weeks or wan­ders for nine months — and almost no one writes that para­graph before they need it.

An invest­ment the­sis is not a memo, not a pitch, not a deck. It is the 35-to-40-word argu­ment that says, in one breath, why this com­pa­ny will return cap­i­tal at this stage, in this mar­ket, with this team, against this set of risks. Every oth­er arti­fact a fund pro­duces — the memo, the term sheet, the part­ner­ship-meet­ing debate — col­laps­es back to that para­graph. The founders who close the clean­est rounds are the ones who wrote that para­graph them­selves before the first investor email went out.

This guide walks through what an invest­ment the­sis real­ly is, the three fla­vors of the­sis you will encounter (fund-lev­el, deal-lev­el, and founder-lev­el), a 40-word tem­plate you can fill in tonight, five ful­ly worked SaaS sam­ples at dif­fer­ent stages, the five ways the­ses fail under part­ner­ship-meet­ing pres­sure, and a $5M annu­al recur­ring rev­enue (ARR) exam­ple that shows the math.

The read­er who will get the most out of the next 20 min­utes is a SaaS chief exec­u­tive offi­cer (CEO) some­where between $3M and $20M ARR, get­ting ready to raise or sell, who wants to stop being a pas­sive sub­ject in some­one else’s argu­ment about your busi­ness.

1. What an Investment Thesis Actually Is

An invest­ment the­sis is a writ­ten, fal­si­fi­able argu­ment that explains why a spe­cif­ic invest­ment (a fund’s strat­e­gy, a sin­gle deal, or a founder’s own raise) will pro­duce returns above an alter­na­tive use of the same cap­i­tal.

Three words in that sen­tence do all the work.

Argu­ment. Not a descrip­tion, not a sum­ma­ry. A the­sis takes a posi­tion. “We invest in B2B soft­ware” is not a the­sis. “We invest in ver­ti­cal B2B soft­ware-as-a-ser­vice (SaaS) for high­ly reg­u­lat­ed mid-mar­ket indus­tries where the buy­er is a non-tech­ni­cal oper­a­tor, because incum­bent on-premise ven­dors have failed to ship mod­ern user expe­ri­ences and switch­ing cost is cre­at­ing win­ner-take-most dynam­ics” is a the­sis.

Fal­si­fi­able. A the­sis names con­di­tions that could prove it wrong. If your the­sis is “AI will be big,” noth­ing could fal­si­fy it, and there­fore noth­ing fol­lows from it. If your the­sis is “ver­ti­cal SaaS for prop­er­ty man­agers above 5,000 units can sus­tain 130%-plus net rev­enue reten­tion because the seat count grows with the cus­tomer’s port­fo­lio,” you have just named a num­ber that can be checked.

Above an alter­na­tive. A the­sis exists rel­a­tive to oth­er uses of the same dol­lar. A ven­ture cap­i­tal (VC) fund’s the­sis com­petes against every oth­er fund a lim­it­ed part­ner (LP) could back. A deal-lev­el the­sis com­petes against the oth­er deals in the part­ner’s pipeline. A founder’s the­sis com­petes against the pub­lic mar­kets and against an LP’s oth­er pri­vate com­mit­ments. If your the­sis does not implic­it­ly answer “why this and not that,” it is not yet a the­sis.

This is why so many invest­ment the­sis sam­ples on the open web read flat. They list focus areas — “stage, geog­ra­phy, sec­tor” — with­out ever stat­ing why those choic­es win. The list is the easy half. The argu­ment is the hard half. The whole point of this arti­cle is to write the argu­ment.

2. The Three Flavors of Investment Thesis

The phrase “invest­ment the­sis” gets used for three dif­fer­ent doc­u­ments, writ­ten by three dif­fer­ent peo­ple, for three dif­fer­ent audi­ences. Con­fus­ing them is the most com­mon mis­take on the web, and it is why most the­sis sam­ples are not use­ful when you sit down to write your own.

ThesisAuthorAudienceLengthPurpose
Fund-level thesisVC general partnerLimited partners (the fund's investors)35–40 words core, 2–4 pages fullArgue why this fund's strategy beats other funds for the same LP dollar
Deal-level thesisSponsoring partner inside the fundThe fund's investment committee1–2 paragraphs in a longer investment memoArgue why this specific company is worth the fund's check at this price
Founder-level thesisThe CEO (you)Yourself, your board, your future investors1 page (you carry this in your head)Force the discipline of pre-deciding the argument others will make about your company

The fund-lev­el the­sis is the one Car­ta, Sequoia, and Tiny Seed have pub­lished. It is strate­gic and pub­lic-ish. The deal-lev­el the­sis is the one buried in the memo your investor writes — short, sharp, deal-spe­cif­ic. The founder-lev­el the­sis is the one almost no founder writes, and it is the high­est-lever­age 90 min­utes you can spend before a raise.

Why this mat­ters for the search. If you typed “invest­ment the­sis sam­ple” look­ing for the fund-lev­el tem­plate, jump to Sec­tion 4. If you are a founder prepar­ing to raise, the founder-lev­el the­sis in Sec­tion 7 is the one that actu­al­ly moves your clos­ing odds. Read both.

3. The Anatomy of a 40-Word Thesis

The mechan­ic behind a tight invest­ment the­sis is the same mechan­ic behind a tight ele­va­tor pitch — a fixed gram­mar with five slots. Fill the slots truth­ful­ly, in this order, and you have a the­sis. Skip any slot and you have a hope.

[Fund or com­pa­ny name] is [stage] invest­ing [check size band] into [sec­tor or sub-sec­tor] [geog­ra­phy] that [shared mechan­ic — the why-this-will-work-pat­tern], lever­ag­ing [the secret sauce — the oper­a­tor’s unfair advan­tage].

That is the gram­mar. Five slots, rough­ly 35–40 words. Here is the rule for each.

  1. Stage. Pre-seed, seed, Series A, Series B, growth, buy­out. Stage dri­ves every oth­er choice. A seed fund and a Series B fund oper­at­ing in the same sec­tor are run­ning com­plete­ly dif­fer­ent busi­ness­es.
  2. Check size. A range, not a point. “$500K to $2M ini­tial” is a the­sis. “We are flex­i­ble” is the absence of a the­sis.
  3. Sec­tor or sub-sec­tor. The nar­row­er the bet­ter. “B2B SaaS” is not a the­sis. “Ver­ti­cal B2B SaaS for com­pli­ance-heavy mid-mar­ket indus­tries” is clos­er. “Work­flow soft­ware for the back office of region­al hos­pi­tals between 200 and 800 beds” is a the­sis.
  4. Geog­ra­phy. A real geo­graph­ic con­straint, even if the con­straint is glob­al. “North Amer­i­ca” is a the­sis. “Any­where” is a mar­ket­ing line.
  5. Shared mechan­ic. This is the slot that sep­a­rates real the­ses from Pow­er­Point dec­o­ra­tion. It is the why-this-pat­tern-wins-now claim: a struc­tur­al mar­ket change, a buy­er-behav­ior shift, a tech­nol­o­gy unlock. “Cloud is replac­ing on-premise soft­ware” was the shared mechan­ic of 2010-era SaaS the­ses. “Gen­er­a­tive arti­fi­cial intel­li­gence (AI) is col­laps­ing the cost of pro­duc­ing per­son­al­ized B2B con­tent” might be a 2026-era shared mechan­ic. With­out this slot, your the­sis is just a list of focus areas.
  6. Secret sauce. The oper­a­tor’s track record trans­lat­ed into why this man­ag­er (or this founder) will win this game. Stat­ed as evi­dence, not as adjec­tive. “Three of our part­ners scaled SaaS com­pa­nies from $5M to $100M+ ARR before becom­ing investors” is evi­dence. “Deep oper­a­tor net­work” is not.

Three rules apply to every slot.

  • No invent­ed sec­tors. “Lazy tech,” “inno­va­tion orig­i­na­tion,” “next-gen unstruc­tured work­flows” — part­ners and lim­it­ed part­ners read these as evi­dence the writer does not actu­al­ly know the mar­ket. Use sec­tor names a Bloomberg ter­mi­nal would rec­og­nize.
  • No range padding. A the­sis that tar­gets “seed through Series C” across “North Amer­i­ca and select inter­na­tion­al” is not a the­sis; it is an unde­cid­ed port­fo­lio.
  • No more than 40 words. The 40-word ceil­ing forces the writer to pick. Most ear­ly the­sis drafts run 80+ words. The edit­ing pass is where the think­ing hap­pens.

The gram­mar at a glance — six inputs col­laps­ing into one para­graph that then dri­ves every down­stream arti­fact:

Inputs (the six slots)OutputWhat the 40 words feed
Stage, check size, sector, geography, shared mechanic, secret sauceThe 35-to-40-word thesis paragraphLP fund deck, deal-level investment memo, term sheet negotiation

If any slot is emp­ty, the para­graph leaks. If the para­graph leaks, every down­stream arti­fact (deck, memo, term sheet lan­guage) inher­its the leak. Fix the slots first; every­thing else fol­lows.

Worked Investment Thesis Samples — Five identical translucent geometric blocks arranged in a ro

4. Five Worked Investment Thesis Samples

Read­ing abstrac­tions is one thing; read­ing a fin­ished the­sis line-by-line is anoth­er. Below are five sam­ples — three fund-lev­el, two deal-lev­el — sized to the 40-word tar­get. Each is fol­lowed by a short read-through of which slot is doing what and which slot is doing too much or too lit­tle.

Note on the num­bers and named firms. The five sam­ples below are illus­tra­tive com­po­si­tions writ­ten to demon­strate the struc­ture of a cred­i­ble invest­ment the­sis. They draw on the shape of the­ses pub­lished by real firms (Besse­mer, Insight Part­ners, Point Nine, Tiny Seed, Open­View and oth­ers) but should not be read as offi­cial state­ments of any actu­al fund’s strat­e­gy. Fund man­dates, check sizes, and sec­tor focus­es change over time — always ver­i­fy any spe­cif­ic fir­m’s cur­rent the­sis against their own mate­ri­als before rely­ing on it.

Sample 1 — Seed SaaS Fund (B2B, North America)

North­beam Ven­tures is a $40M seed-stage fund writ­ing $500K–$1.5M ini­tial checks into U.S. and Cana­di­an busi­ness-to-busi­ness soft­ware-as-a-ser­vice com­pa­nies serv­ing reg­u­lat­ed mid-mar­ket indus­tries, back­ing the first wave of ver­ti­cal AI agents replac­ing man­u­al back-office work­flows, led by three for­mer ver­ti­cal-SaaS oper­a­tors with com­bined $400M+ in pri­or exits.

Word count: 53. Already over the 40-word ceil­ing; in a real sub­mis­sion, the secret-sauce phrase would tight­en to “led by oper­a­tors with pri­or $400M+ ver­ti­cal-SaaS exits” (~38 words total). But the slots are clean: stage (seed), check size ($500K–$1.5M), sec­tor (ver­ti­cal B2B SaaS, reg­u­lat­ed mid-mar­ket), geog­ra­phy (U.S. and Cana­da), shared mechan­ic (ver­ti­cal AI agents replac­ing man­u­al back-office work­flows), secret sauce (oper­a­tor exits). A lim­it­ed part­ner can decide whether to take a first meet­ing from this para­graph alone.

Sample 2 — Series A Fund (Global, Horizontal SaaS)

Lat­i­tude Cap­i­tal is a Series A fund writ­ing $5M–$12M ini­tial checks into hor­i­zon­tal B2B SaaS com­pa­nies any­where glob­al­ly with $1M–$5M ARR, 100%+ year-over-year growth, and net rev­enue reten­tion above 115%, part­nered with the founders of three com­pa­nies that exit­ed above $1B.

Word count: 41. The “shared mechan­ic” slot here is implic­it — the the­sis backs com­pa­nies that have already proven prod­uct-mar­ket fit at the ear­ly-rev­enue stage, bet­ting that those com­pa­nies will com­pound. That is a legit­i­mate mechan­ic, but a sharp­er ver­sion would name why this stage right now (e.g., “as infla­tion in seed-stage round sizes push­es the sur­vivor bar at Series A high­er than ever”). Still, the slots are filled and the LP can act on it.

Sample 3 — Bootstrapper-First Seed Fund

Wild­flower Fund is a $20M seed fund writ­ing $120K–$300K checks into boot­strapped B2B SaaS com­pa­nies between $200K and $1M ARR glob­al­ly, believ­ing inde­pen­dent SaaS out­comes are pow­er-law dis­trib­uted and that a broad port­fo­lio with no fol­low-on pres­sure pro­duces top-decile fund returns.

Word count: 42. This is the clean­est of the three because the shared mechan­ic (“pow­er-law dis­trib­uted out­comes, broad port­fo­lio, no fol­low-on pres­sure”) is not a sec­tor claim — it is a strat­e­gy claim. Com­pare to the pub­lic Tiny Seed the­sis (which inspired this com­po­si­tion) — the explic­it nam­ing of the strat­e­gy is the argu­ment. The­ses that argue strat­e­gy beat the­ses that just argue sec­tor when the sec­tor itself is crowd­ed.

Sample 4 — Deal-Level Thesis (Series A SaaS)

We rec­om­mend lead­ing Vec­tor Ana­lyt­ics’ $12M Series A at $48M post-mon­ey, bet­ting that the buy­er pat­tern we are see­ing in mid-mar­ket pro­cure­ment — non-tech­ni­cal oper­a­tors reject­ing lega­cy on-premise tools the month after a CFO turnover — pro­duces a cat­e­go­ry-defin­ing work­flow ven­dor with 124% net rev­enue reten­tion com­pound­ing into $40M+ ARR in 36 months.

Word count: 56. Long by fund-the­sis stan­dards, but deal-lev­el the­ses are slight­ly more per­mis­sive because they have to name the price. The slots: stage and check size (Series A, $12M), sec­tor (mid-mar­ket pro­cure­ment work­flow SaaS), the shared mechan­ic (“CFO turnover trig­ger­ing on-premise ven­dor dis­place­ment” — spe­cif­ic, fal­si­fi­able, observ­able in cus­tomer inter­views), and the explic­it pro­jec­tion. A part­ner can defend or attack this para­graph in com­mit­tee.

Sample 5 — Deal-Level Thesis (Series B SaaS, Defensive Re-Up)

We rec­om­mend re-upping $8M into Helix Health’s $30M Series B at $180M post-mon­ey to main­tain own­er­ship, bet­ting that the com­pa­ny’s ver­ti­cal posi­tion in region­al hos­pi­tal billing — now the only soft­ware touch­ing 7 of 14 state Medicare-state-admin­is­tra­tion changes — has crossed from “use­ful tool” into “reg­u­lat­ed infra­struc­ture” with 145% net rev­enue reten­tion and 9‑month pay­back on enter­prise expan­sion.

Word count: 60. This is a re-up the­sis (the fund is invest­ing more cap­i­tal into an exist­ing port­fo­lio com­pa­ny), which has dif­fer­ent rhetor­i­cal work to do — the part­ner has to defend why this dol­lar in here beats this dol­lar in a new invest­ment. The slot doing the heavy lift­ing is “reg­u­lat­ed infra­struc­ture” — a cat­e­go­ry tran­si­tion claim that, if true, jus­ti­fies the price; if false, kills the deal. That is exact­ly where part­ner­ship debate should land. Good re-up the­ses sur­face the tran­si­tion claim explic­it­ly so the room can argue it.

The pat­tern across all five: the slots that look gener­ic in a tem­plate (stage, check size, geog­ra­phy) take 10 sec­onds to fill. The slot that earns the meet­ing is the shared mechan­ic — the why this pat­tern wins now claim. The sam­ples that work spend most of their words there.

5. Investment Thesis vs. Investment Memo vs. Pitch Deck

A search for “invest­ment the­sis sam­ple” often sur­faces results for invest­ment mem­os and pitch decks. The three are relat­ed but do dif­fer­ent jobs. Con­fus­ing them is why so many founder-writ­ten the­ses end up read­ing like recy­cled deck copy.

ArtifactLengthAuthorAudienceJob
Investment thesis35–40 words core (1 page with build-out)Fund GP, deal partner, or founderLPs, IC, or the founder themselvesState the argument in one breath
Investment memo5–15 pagesSponsoring partnerThe full partnership (investment committee)Defend the thesis section by section, with data
Pitch deck12–20 slidesFounderInvestors in the first meetingLand the emotional and intellectual hook in 20 minutes

The depen­den­cy runs one direc­tion: deck sup­ports memo sup­ports the­sis. The the­sis is the seed of the memo, and the memo is the seed of the deck. Founders who reverse-engi­neer in the oth­er direc­tion — writ­ing a beau­ti­ful deck first, then back-fill­ing a the­sis — almost always end up with a the­sis that is too broad (because the deck has to flat­ter every pos­si­ble objec­tion) and a memo that is inter­nal­ly incon­sis­tent (because the slots were filled in ser­vice of slide nar­ra­tive rather than truth­ful argu­ment).

If you are about to raise, write the the­sis first. Then write the memo. Then design the deck. The work com­pounds in that order.

The com­pan­ion piece on the invest­ment memo walks through the sev­en sec­tions every memo con­tains and a $5M ARR worked exam­ple. Read­ing both togeth­er gives you the full stack: the one-para­graph argu­ment plus the full doc­u­ment that defends it.

Industry-Standard Investment Thesis Templates — Layered transparent sheets of paper stacked at slight offset

6. Industry-Standard Templates and Where They Come From

The 40-word gram­mar in Sec­tion 3 is not orig­i­nal to this arti­cle. Vari­a­tions of it have been pub­lished by VC Lab, Car­ta, NfX, and Recast Cap­i­tal. The rea­son the same five slots keep sur­fac­ing is that the part­ner­ship-meet­ing deci­sion is the same every­where: a room­ful of part­ners has to allo­cate a finite check across com­pet­ing oppor­tu­ni­ties, in a finite amount of time, against part­ner-spe­cif­ic rep­u­ta­tion costs. The five slots are the min­i­mum infor­ma­tion set the room needs to vote.

If you want to com­pare tem­plates, the VC Lab tem­plate is the most cit­ed, and the Tiny Seed the­sis is a strong exam­ple of the strat­e­gy-as-mechan­ic style demon­strat­ed in Sam­ple 3 above. Read both. The dif­fer­ences across them are sur­face dif­fer­ences — the under­ly­ing gram­mar is the same.

What changes year over year is the shared mechan­ic slot. In 2010, “cloud is replac­ing on-premise” was the dom­i­nant mechan­ic. In 2015, “the iPhone has trained every con­sumer to expect soft­ware-qual­i­ty UX in enter­prise tools” became the mechan­ic. In 2020, “remote work has struc­tural­ly raised will­ing­ness to pay for col­lab­o­ra­tion soft­ware.” In 2026, gen­er­a­tive AI’s effect on white-col­lar work­flow cost is being claimed as the mechan­ic across dozens of new funds — most of those claims will turn out to be too broad, and the funds whose mechan­ic claims are tight­est (spe­cif­ic indus­try, spe­cif­ic work­flow, spe­cif­ic buy­er) will out­per­form.

The les­son for any founder writ­ing their own the­sis: choose your mechan­ic as if you have to defend it for ten years. Gener­ic mechan­ics (“AI is trans­form­ing every­thing”) gen­er­ate gener­ic returns. Spe­cif­ic mechan­ics (“com­pli­ance-dri­ven doc­u­ment review in region­al law firms is mov­ing from junior asso­ciates to AI agents because the super­vis­ing-part­ner work­flow is already dig­i­tized”) gen­er­ate fund­able con­vic­tion.

Founder's Investment Thesis — A solitary antique compass laid on a dark wood surface lit b

7. The Founder’s Investment Thesis: How to Write Yours

The founder-lev­el the­sis is the high­est-return writ­ing you will do before a raise, and almost no founder does it. The rea­son is struc­tur­al: the act of writ­ing the the­sis forces you to admit, in your own words, what you are bet­ting your own time on — and that admis­sion is uncom­fort­able enough that most founders skip it and let investors do the writ­ing for them. The cost shows up in val­u­a­tion, in round speed, and in term sheet qual­i­ty.

The exer­cise takes 90 min­utes, twice. Once two weeks before you start fundrais­ing, and once the night before your first pitch.

Step 1 — Write the 40-Word Argument for Your Own Company

Use the same five-slot gram­mar from Sec­tion 3, but with you as the sub­ject:

[Com­pa­ny name] is a [stage of com­pa­ny] build­ing [spe­cif­ic prod­uct / cat­e­go­ry] for [spe­cif­ic buy­er in spe­cif­ic geog­ra­phy], bet­ting that [the shared mechan­ic — what is chang­ing in your mar­ket right now], lever­ag­ing [your unfair advan­tage as oper­a­tors].

Write it. Stop at 40 words. The dis­ci­pline is the point.

Step 2 — Stress-Test the Mechanic Slot

The mechan­ic slot is where founder the­ses fail. Ask three ques­tions of yours:

  1. What evi­dence would con­vince a skep­tic that the mechan­ic is real? Cus­tomer inter­views? Pipeline data? A spe­cif­ic reg­u­la­to­ry change? A spe­cif­ic com­peti­tor’s bank­rupt­cy? If you can­not name evi­dence, the mechan­ic is a guess.
  2. *What evi­dence would con­vince you the mechan­ic is wrong?* This is the fal­si­fi­a­bil­i­ty test. If noth­ing could change your mind, the mechan­ic is not a claim — it is faith.
  3. How long does the mechan­ic stay true? If the mechan­ic is true for 18 months, you are pitch­ing a win­dow. If the mechan­ic is true for 10 years, you are pitch­ing a cat­e­go­ry. Investors price the two very dif­fer­ent­ly.

Step 3 — Show the Thesis to Five Customers Without Calling It a Thesis

Read the para­graph (the mechan­ic slot in par­tic­u­lar) to five cus­tomers as if it were your descrip­tion of why you start­ed the com­pa­ny. Watch their faces. The cus­tomers who lean for­ward are the ones who already feel the mechan­ic in their day-to-day. If five out of five nod polite­ly with­out lean­ing for­ward, your mechan­ic is not real to your buy­ers yet — fix the mechan­ic before you fix the deck.

Step 4 — Surface the Three Risks That, If True, Kill the Thesis

Write the three things that, if true, would fal­si­fy the argu­ment. Cus­tomer con­cen­tra­tion. A pend­ing reg­u­la­to­ry change. A well-fund­ed incum­bent with a free ver­sion. These are the three risks your investor’s memo will list — your job is to list them first, and to attach a one-para­graph mit­i­ga­tion to each.

Step 5 — Lock the Thesis, Then Build Everything Else From It

Every dili­gence answer, every ref­er­ence call, every term sheet nego­ti­a­tion should rein­force the same the­sis. Founders who change the the­sis across con­ver­sa­tions get marked down on what part­ners polite­ly call “sto­ry drift” — a euphemism for the founder has not fig­ured out what they actu­al­ly believe. The way to avoid drift is to lock the 40-word para­graph and refer back to it before every exter­nal con­ver­sa­tion.

The founder who com­pletes these five steps rais­es faster, at high­er prices, with clean­er term sheets. The founder who skips them walks into part­ner­ship meet­ings being sum­ma­rized by peo­ple they have known for three weeks.

8. A Worked Example: $5M ARR SaaS Founder Thesis

Con­crete is more use­ful than abstract. Here is the founder-lev­el the­sis for a hypo­thet­i­cal $5M ARR Series A SaaS — the same fic­tion­al com­pa­ny (Vec­tor Ana­lyt­ics) used in the com­pan­ion invest­ment memo arti­cle. The num­bers in this sec­tion are illus­tra­tive and reflect typ­i­cal Series A SaaS bench­marks; they are includ­ed to show the shape of a cred­i­ble the­sis, not to imply that any spe­cif­ic num­ber is cur­rent-mar­ket for your deal. Always bench­mark against cur­rent data before lock­ing your own the­sis.

The 40-word the­sis (writ­ten by the CEO, two weeks before the raise):

Vec­tor Ana­lyt­ics is a $5M ARR Series A work­flow ana­lyt­ics plat­form for North Amer­i­can mid-mar­ket pro­cure­ment teams, bet­ting that non-tech­ni­cal oper­a­tors are aban­don­ing lega­cy on-premise tools with­in 90 days of a CFO turnover, lever­ag­ing our two founders’ 11 com­bined years inside For­tune 1000 pro­cure­ment.

Word count: 41. Slots:

  • Stage: Series A
  • Check size: implic­it ($12M raise — stat­ed in the memo, not the the­sis)
  • Sec­tor: work­flow ana­lyt­ics SaaS, mid-mar­ket pro­cure­ment
  • Geog­ra­phy: North Amer­i­ca
  • Shared mechan­ic: CFO turnover trig­gers on-premise ven­dor dis­place­ment with­in 90 days
  • Secret sauce: 11 com­bined years inside For­tune 1000 pro­cure­ment

The mechan­ic slot here is doing the heavy lift­ing. It is spe­cif­ic, fal­si­fi­able (you can check whether new cus­tomers cor­re­late with CFO tran­si­tions), and time­ly (the post-pan­dem­ic CFO turnover wave is observ­able in S&P 500 dis­clo­sure data).

The Math That Backs the Thesis

A the­sis with­out num­bers is a wish. Vec­tor Ana­lyt­ic­s’s the­sis has to be defen­si­ble against the same nine met­rics every Series A SaaS investor will check (the full list is in the invest­ment memo arti­cle). The trac­tion snap­shot:

MetricValueSeries A BenchmarkRead
Annual recurring revenue (ARR)$5.2M$3M–$10MIn range
Year-over-year ARR growth90%80%–120%In range
Net revenue retention (NRR)124%110%+Strong
Gross revenue retention (GRR)91%90%+In range
LTV to CAC ratio4.6 : 13:1+Strong
CAC payback (months, gross profit)28.818–30In range
Gross margin76%75%+In range
Magic Number (trailing 12 months)0.90.7+Strong
Burn multiple1.4<2Strong

The the­sis is con­sis­tent with these num­bers. A the­sis claim­ing “cat­e­go­ry-defin­ing” with 60% growth and 95% NRR would be incon­sis­tent — part­ners would mark it down. The slot-and-num­ber con­sis­ten­cy check is where weak the­ses die.

LTV to CAC — The Math Shown

Show the math behind any ratio in a the­sis. If the the­sis cites 4.6 : 1, show how:

  • Aver­age rev­enue per account (ARPA): $52,000 per year
  • Gross mar­gin: 76%
  • Annu­al gross prof­it per account: $52,000 × 0.76 = $39,520
  • Gross rev­enue reten­tion: 91% → implied annu­al cus­tomer churn ≈ 9%
  • Cus­tomer life­time (years): 1 ÷ 0.09 ≈ 11.1 years
  • Cus­tomer life­time val­ue (LTV): $39,520 × 11.1 ≈ $438,672
  • Blend­ed cus­tomer acqui­si­tion cost (CAC): $95,000
  • LTV:CAC ratio: $438,672 ÷ $95,000 ≈ 4.62 : 1

A part­ner read­ing this can red­erive the ratio in 90 sec­onds. That is the point. The­ses (and mem­os) that show the math get believed. The­ses that just claim “LTV:CAC of 4.6” with­out the inputs get ques­tioned — and a the­sis under ques­tion­ing los­es momen­tum the room can­not recov­er.

CAC Payback — The Math Shown

  • Blend­ed CAC: $95,000
  • Month­ly gross prof­it per account: $39,520 ÷ 12 ≈ $3,293
  • CAC pay­back (months, gross prof­it basis): $95,000 ÷ $3,293 ≈ 28.8 months

A read­er will notice this is the strict gross-prof­it ver­sion — CAC divid­ed by month­ly gross prof­it per account, with no growth-in-peri­od or con­tri­bu­tion-mar­gin adjust­ment. Some dash­boards report pay­back dif­fer­ent­ly, and a the­sis cit­ing 14- or 18-month pay­back should always foot­note the for­mu­la. The num­ber itself mat­ters less than the con­sis­ten­cy between for­mu­la and inputs — investors who notice mis­match­es usu­al­ly con­clude the founder does not know which ver­sion their finance team uses, which becomes a CFO-qual­i­ty flag, not just a met­ric flag.

The rea­son to show the math inside a the­sis worked-exam­ple arti­cle is that the math is what makes the the­sis fal­si­fi­able. A read­er can argue with the gross mar­gin assump­tion (is 76% sus­tain­able?), with the churn assump­tion (is 9% the right base rate?), and with the CAC assump­tion (is $95,000 blend­ed or new-logo-only?). A the­sis a read­er can argue with is a the­sis. A the­sis nobody can argue with is a slo­gan.

Investment Thesis Failure Patterns — A cracked sheet of dark ice across the lower half of the fra

9. The Five Ways an Investment Thesis Fails

After enough part­ner­ship-meet­ing cycles, the same the­sis fail­ures recur. Avoid them.

Failure 1 — The Mechanic Is a Trend, Not a Mechanism

“AI is chang­ing every­thing” is a trend state­ment. “Gen­er­a­tive AI is col­laps­ing the cost of pro­duc­ing per­son­al­ized B2B mar­ket­ing con­tent by 10x, which is struc­tural­ly com­press­ing cre­ative-agency gross mar­gins” is a mech­a­nism. The first is a the­sis the part­ner can­not defend in com­mit­tee because the LP just heard six oth­er founders say it. The sec­ond is a the­sis that names a spe­cif­ic buy­er-eco­nom­ics change and there­fore gen­er­ates spe­cif­ic port­fo­lio bets.

Failure 2 — The Slots Are Filled but the Argument Is Missing

It is pos­si­ble to fill all six slots — stage, check size, sec­tor, geog­ra­phy, mechan­ic, secret sauce — and still pro­duce a the­sis with no argu­ment in it. The test: read your the­sis to a stranger and ask, “why does this strat­e­gy beat the alter­na­tive?” If you can­not answer in one sen­tence, you have pro­duced a descrip­tion, not a the­sis.

Failure 3 — The Thesis Is Internally Inconsistent With the Numbers

A the­sis claim­ing “we back cat­e­go­ry-defin­ing com­pa­nies” paired with a memo show­ing 50% growth and 100% NRR is incon­sis­tent. Cat­e­go­ry-defin­ing com­pa­nies show 100%+ growth and 130%+ NRR. The inter­nal-con­sis­ten­cy check is what kills weak the­ses in the part­ner­ship meet­ing — and part­ners catch it instant­ly because they spend their days match­ing argu­ments to num­bers.

Failure 4 — The Thesis Is Too Broad to Be Falsified

“We invest in great founders build­ing impor­tant com­pa­nies.” This is the most com­mon fail­ure on the open web. It can­not be fal­si­fied, which means it can­not be test­ed, which means it gen­er­ates no actu­al deci­sions. Re-write the slots until the the­sis names con­di­tions under which it would be proven wrong.

Failure 5 — The Thesis Drifts Across Conversations

A founder-lev­el the­sis you change between the first pitch and the part­ner meet­ing reads to investors as con­fu­sion. Lock the 40-word para­graph two weeks before the raise. Car­ry it on a card in your wal­let if you have to. Every con­ver­sa­tion rein­forces the same para­graph; every dili­gence answer con­nects back to the same mechan­ic; every ref­er­ence call is briefed with the same shared mechan­ic in mind. Sto­ry drift is the silent killer of rounds at oth­er­wise excel­lent com­pa­nies.

10. Frequently Asked Questions

What is an investment thesis sample, in plain English?

An invest­ment the­sis sam­ple is a worked exam­ple of the 35-to-40-word argu­ment that explains why a spe­cif­ic invest­ment strat­e­gy (a fund’s, a deal part­ner’s, or a founder’s) will pro­duce returns above an alter­na­tive use of the same cap­i­tal. The five slots are stage, check size, sec­tor, geog­ra­phy, and the shared mechan­ic — the spe­cif­ic mar­ket change that makes this strat­e­gy work right now. Sec­tion 4 of this arti­cle gives five sam­ples.

How long should an investment thesis be?

The core argu­ment should be 35–40 words. A full the­sis doc­u­ment — fund-lev­el — runs 2–4 pages with the build-out (team, track record, port­fo­lio con­struc­tion mod­el). A deal-lev­el the­sis fits in 1–2 para­graphs inside a longer invest­ment memo. A founder-lev­el the­sis lives on a one-page doc­u­ment the CEO car­ries in their head. The 40-word ceil­ing is on the argu­ment, not on the sup­port­ing mate­r­i­al.

How is an investment thesis different from an investment memo?

An invest­ment the­sis is the one-para­graph argu­ment. An invest­ment memo is the 5‑to-15-page doc­u­ment that defends the the­sis sec­tion by sec­tion with data. The the­sis is the seed; the memo is the tree. The pitch deck is the leaves the founder shows the world. Con­fus­ing the three is the most com­mon writ­ing mis­take on the open web.

What is the difference between a fund-level and a founder-level investment thesis?

A fund-lev­el the­sis is writ­ten by a VC gen­er­al part­ner for the fund’s lim­it­ed part­ners (LPs) — it argues why this fund’s strat­e­gy will beat alter­na­tives. A founder-lev­el the­sis is writ­ten by the CEO for them­selves — it argues why this com­pa­ny at this stage in this mar­ket is worth the founder’s next five years and an investor’s next $10M+. Both use the same five-slot gram­mar. The audi­ence and the implied “alter­na­tive use of cap­i­tal” dif­fer.

Do I need to write an investment thesis if I am bootstrapping?

Yes — although you call it some­thing else. A boot­strapped CEO writ­ing a the­sis is forc­ing them­selves to artic­u­late what they are bet­ting their own time on, against alter­na­tives like a full-time job at a pub­lic SaaS com­pa­ny or run­ning a dif­fer­ent busi­ness. The mechan­ic slot in par­tic­u­lar — what is chang­ing in my mar­ket right now that makes my next five years valu­able? — is the dis­ci­pline that sep­a­rates boot­strap­pers who com­pound cap­i­tal from boot­strap­pers who run lifestyle busi­ness­es with­out real­iz­ing it.

What is the most common mistake in investment thesis samples I find online?

The most com­mon mis­take is fill­ing the slots with­out writ­ing the argu­ment. A the­sis that names a stage, check size, sec­tor, and geog­ra­phy but does not name a shared mechan­ic — the spe­cif­ic mar­ket change that makes the strat­e­gy win — is a descrip­tion, not a the­sis. The fix is to spend 80% of the writ­ing time on the mechan­ic slot and 20% on every­thing else.

Can a generative AI tool write my investment thesis?

A gen­er­a­tive AI tool can scaf­fold the slots and pro­pose mechan­ic can­di­dates, but it can­not write a defen­si­ble mechan­ic because the defense requires evi­dence the AI does not have — your cus­tomer inter­view data, your pipeline com­po­si­tion, your cat­e­go­ry-spe­cif­ic knowl­edge of which incum­bent is fal­ter­ing. Use AI to draft 10 can­di­date mechan­ic state­ments, then test each one against your own cus­tomer data. The the­sis stays yours.

11. The Takeaway

An invest­ment the­sis is not a tem­plate to fill in. It is a 40-word argu­ment that has to sur­vive a part­ner­ship meet­ing, a board review, or the founder’s own quar­ter­ly self-hon­esty check. The five slots are the easy part. The mechan­ic slot — the why this pat­tern wins now claim — is the part that earns or los­es the round.

Write the 40-word ver­sion of your the­sis tonight. Stress-test the mechan­ic with five cus­tomers tomor­row. Then watch how every oth­er arti­fact you pro­duce — the deck, the memo, the term sheet con­ver­sa­tion, the cus­tomer ref­er­ence call — gets faster, sharp­er, and more believ­able because there is one para­graph at the cen­ter of it all.

The founders who do this raise faster, at high­er prices, with clean­er terms. The founders who skip it walk into part­ner­ship meet­ings being sum­ma­rized by peo­ple who have known them for three weeks. Choose accord­ing­ly.

Relat­ed Read­ing:

Facebooktwitterlinkedinmail
author avatar
Vic­tor Cheng
Author of Extreme Rev­enue Growth, Exec­u­tive coach, inde­pen­dent board mem­ber, and investor in SaaS com­pa­nies.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top