How Much Is My Business Worth? 3 SME Valuation Methods

"I heard a competitor sold for several times earnings." "Someone in our industry got two times revenue." These numbers travel fast at trade association dinners and supply-chain lunches, and they almost never come with the conditions attached: was it a share deal or an asset deal? Was the buyer a strategic competitor or a financial investor? Was the price paid at closing, or tied to three years of earn-out targets? Change any one of those conditions and the same company can be worth twice as much, or half.

So most owner-operators who have run a business for twenty or thirty years carry a question they have never actually asked out loud: what is my company actually worth?

This article will not give you a number. Anyone who quotes you a price before reading your financial statements, customer contracts and asset register is not valuing your business, they are guessing. What this article does instead is more useful: it takes the valuation machinery apart so you can see how it works. Three mainstream approaches, what each one is good for, where each one is blind, and the factors that actually push a multiple up or cut it down. You will not walk away with an answer, but you will know what data you are missing, who to talk to, and which three questions to ask when someone finally quotes you a figure.

The timing matters in Taiwan. Taiwan’s Ministry of Economic Affairs reports in the 2025 SME White Paper that the country had more than 1.715 million SMEs in 2024, over 98% of all enterprises, employing roughly 9.194 million people. PwC’s 2025 family business survey found that more than 60% of Taiwanese family businesses have a succession plan under way, with 46% intending to hand over to children or family members and 16% to professional managers. A large amount of ownership is about to change hands, and before it moves, somebody has to know what it is worth.

If you are thinking about the whole succession question, start with the HappyCXO Studio succession and business sale hub. This piece focuses narrowly on valuation mechanics.

Myth number one: there is no single number called "company value"

Direct answer: value is not a number, it is a range that holds only under a stated set of assumptions, for a specific buyer, under specific deal terms. Change any premise and the number changes. So the right question is never "what is my company worth" but "under what assumptions, sold to whom, on what terms, roughly what range."

Professional practice is explicit about this. The International Valuation Standards published by the IVSC group valuation methods into three families — the market approach, the income approach and the cost or asset approach — and require the valuer to define a basis of value and a premise of value before any arithmetic begins. Taiwan has its own formal framework: the Valuation Standards Committee of the Accounting Research and Development Foundation publishes a full set of valuation standards bulletins, including Bulletin No. 15 on valuation approaches and specific valuation methods, issued in March 2023, alongside Bulletin No. 13 on investigation and compliance and No. 14 on non-financial liabilities. Valuation is a discipline with procedures, not an opinion.

So why does one company produce different numbers? Because the basis of value differs. Liquidation value assumes you shut down tomorrow. Going-concern value assumes the business keeps running. Investment value is specific to one buyer — a strategic competitor who can strip out duplicate overhead, share distribution channels and load your capacity with orders he already has will pay more than a purely financial investor. That is not generosity, it is the arithmetic of synergy.

One level deeper sits the trap that catches most owners: share deals and asset deals are not price-comparable. Buying shares means buying the company’s history too, including unresolved tax exposure, labour disputes, environmental liability and guarantees. Buying assets means picking out the plant, equipment, inventory and customer list and leaving the history behind. For the same underlying assets, a share deal is usually priced lower because the buyer discounts for latent risk, while an asset deal may price higher but leave the seller with a heavier tax bill. If the multiple you heard did not specify which structure it was, it carried no information.

McKinsey states the underlying logic cleanly: growth and return on invested capital together drive cash flow, and cash flow is what ultimately drives value. That sentence deserves a second reading from any SME owner. Revenue by itself creates nothing. What creates value is durable cash flow above that revenue, and how much capital you must keep feeding in to produce it. A company doing NT$300 million a year that survives only by continuously buying machines and stockpiling material can be worth less than one doing NT$100 million that needs almost no incremental capital.

One last piece of mental preparation. The purpose of a valuation shapes its output. A valuation done to sell, to split a family estate, to support a bank loan, or to price an employee equity plan can each land on a different figure, and each of them can be correct. When two professionals give you different numbers, do not assume one is incompetent. Ask what basis of value and what premise each one used.

Approach one: the asset approach, and where it goes blind

Direct answer: the asset approach restates every asset at market value, subtracts liabilities, and produces a net asset value. It suits asset-heavy, unprofitable or soon-to-be-liquidated companies, and it usually functions as the floor of value rather than the price a healthy seller should target.

The intuition matches how most manufacturing owners already think. The land, the building, the machines, the inventory, minus the bank debt, equals my net worth. Nothing wrong with the instinct. What is wrong is treating the book number as the market number. Book equity almost never equals the asset approach result, because financial accounting optimises for prudence and verifiability while valuation optimises for present transactional reality.

In practice the adjustments fall into four buckets. Land and buildings come first: many older Taiwanese factories sit on land acquired forty years ago and carried at historical cost, so a revaluation can be a multiple of book — often the single largest source of real value, and often left out of the negotiation entirely. Machinery is second: fully depreciated equipment is not worthless, but a specialised machine still on the books can have almost no secondary market if the industry has moved on. Inventory is third: slow-moving stock, spares older than three years and material bought for one departed customer are worth something very different to a buyer than to your ledger. Receivables are fourth: anything over a year old, or owed by a customer already in trouble, will be written off in full by the buyer.

The first blind spot is that the asset approach barely sees intangibles at all. Twenty years of customer relationships, hard-won certifications, the process parameters in a senior technician’s head, the quality reputation the trade takes for granted — none of it has a home in this method unless you can convert it into something identifiable and transferable: patents, trademarks, documented licensable process, contracted long-term customers. Which is why a profitable, stable business should almost never accept a pure asset-based offer. Conversely, if a buyer opens with asset-approach language only, he is telling you he does not believe your earnings are durable.

The second blind spot is subtler: the asset approach makes people overestimate how easily they could liquidate. Liquidation is a process, not a moment. Specialised machines can take six to twelve months to place, and the buyer is always the one with leverage. Industrial-zone property brings zoning questions and environmental restoration duties. Meanwhile you keep paying wages, rent and interest. Real net liquidation proceeds are usually meaningfully below the paper figure.

So when does the asset approach lead? Three situations. When the company loses money or earns erratically, so an income-based number carries no credibility. When the assets essentially are the business, as with a traditional factory whose land and buildings dominate. And for non-transactional purposes: estate division, inheritance, bank collateral assessment, court proceedings. Otherwise its correct role in an SME sale is simple: it sets a floor so you do not sell too cheaply, but a good price has to be earned by the other two approaches.

Approach two: market comparables, and the trap inside them

Direct answer: the market approach infers your value from what comparable companies actually sold for. It is the most intuitive method, and in the Taiwanese SME world its fatal problem is that comparable data barely exists, or exists but cannot be verified.

Start with why it matters. The market approach reflects real transactions rather than model assumptions, which makes it the hardest method to argue away. McKinsey also cautions that multiples express value rather than create it, and that you should first understand where the value is being created, with the multiple as a translation of that into a comparable figure. The multiple is an output, not a cause.

Here is the problem: there is no public, verifiable statistic for Taiwanese SME transaction multiples. Private-company deals in Taiwan close by private agreement, with no disclosure obligation and no regulator or trade body publishing a median. So when someone tells you that machinery businesses in Taiwan currently sell at some specific multiple, you are entitled to ask for the source, the sample size and the period. If the answer is not forthcoming, it is a rumour, not data.

Public statistics do exist for the United States. BizBuySell publishes a quarterly Insight Report on small-business transactions, and its Q1 2026 report put the median sale price at US$350,000, median cash flow at about US$165,000, median revenue at roughly US$713,000, and the average cash flow multiple at 2.7x. Note three things: this is US data, it is based on cash flow or seller discretionary earnings rather than EBITDA, and the sample skews to very small main-street businesses. Applying 2.7x to a Taiwanese precision machining shop would be a serious misuse.

One tier up, the IBBA and M&A Source survey member-completed deals each quarter in the Market Pulse report. The Q1 2026 edition polled 300 business brokers and M&A advisors covering 203 completed transactions, and defines main street as US$0 to US$2 million and the lower middle market as US$2 million to US$50 million. Two findings travel well to Taiwan: 83% of deals above US$5 million attracted at least three offers and 18% attracted ten or more bids, and on the question of AI, 67% of advisors saw no material valuation impact yet, 12% saw upside, 3% saw downside and 15% called it too early. The first says scale creates competitive tension among buyers. The second says the market is still forming a view on AI. Both are US data, and should be labelled as such in any conversation.

So how should a Taiwanese seller actually use the market approach? Three practical moves. First, use listed peer EV/EBITDA as an upper reference and then discount honestly — public companies have liquidity, governance and audited statements, and an unlisted SME lacks all three. There is no formula for the discount; it is a negotiation. Second, treat every multiple you hear as a lead to be verified, and ask four questions: share deal or asset deal, is the denominator EBITDA or net profit or revenue, was there an earn-out or instalment structure, and how long did the seller stay on. Third, accept that in a market without comparables, the market approach is a sanity check rather than a pricing tool. Its job is to tell you whether your expectation is wildly off, not to give you the answer.

Approach three: the income approach and EBITDA multiples

Direct answer: the income approach asks how much cash this business can keep generating and what that stream is worth today. It is what professional buyers actually calculate, and its common SME shorthand is adjusted EBITDA multiplied by a multiple.

Definitions first. EBITDA strips out interest, tax, depreciation and amortisation in order to separate whether the business itself earns money from how the owner has arranged financing and tax. Buyers care about the former, because they will rearrange the latter their own way. Underneath sits discounting: money today is worth more than the same money later, because today’s money can be reinvested and because inflation erodes future purchasing power. Harvard Business Review’s primer on net present value explains that time value clearly. The whole income approach is simply future cash pulled back to today at a rate that reflects risk.

Full discounted cash flow modelling is rare in SME deals because the forecasts are not reliable enough. Multiples take its place, and the important part is not the multiple, it is the denominator. Adjusted or normalised EBITDA — sometimes expressed as seller discretionary earnings — restores everything that does not belong to ongoing operations: excess owner and family compensation, personal vehicles and insurance carried by the company, one-off litigation or relocation costs, rental income from non-operating assets, and related-party transactions priced away from market. That schedule frequently determines more than half the final price, and it depends almost entirely on the quality of your books.

Which leads to the most painful subject for Taiwanese SMEs: two sets of books. Many companies have spent years reporting far less profit than they earn, for tax reasons. The owner knows the real number. But a buyer can only buy what he can verify. When you tell a buyer that you really make NT$20 million while the books show NT$5 million, his rational responses are, in order: discount for verification cost and retrospective tax exposure, wonder what else has not been disclosed, and if due diligence cannot substantiate the gap, price off the reported figure. That is not the buyer being harsh; it is what any rational party does under information asymmetry. Bookkeeping transparency is not a moral question, it is a valuation question.

How is the multiple itself set? A multiple is the mirror image of a discount rate and a growth rate. Lower risk, higher predictability and stronger growth produce a higher multiple. Two companies earning identical EBITDA can therefore be worth very different amounts, and the difference lies entirely in the quality behind the number — which is what the next section is about.

One more reference point, and again it is overseas data. Stanford Graduate School of Business tracks search funds, a model in which a professional manager raises capital, buys one small company and runs it personally. Its 2026 research reports more than 850 core search funds tracked since 1996, an aggregate IRR of 33.9% and a 4.75x return multiple, with a median purchase price of US$16 million for 2024–25 acquisitions and roughly 20 months from raise to acquisition. The relevance to a Taiwanese seller is not the multiple; it is the shape of the buyer universe. An entire professional buyer class exists whose business model is acquiring sound companies whose owners want to retire without a successor. If your company is in good shape and your data is complete, the buyer pool is larger than you assume.

Comparison table: when each approach applies, what data it needs, where it goes wrong

Direct answer: the three approaches do not replace one another, they cross-check one another. Good practice runs all three and looks at where they converge. If the three numbers diverge wildly, that divergence is itself the most valuable diagnostic you will get.

Read that table right to left. The income approach sets your ceiling, the market approach sets your plausibility, the asset approach sets your floor. When the income result sits clearly above the asset result, the company has created value beyond its tangible assets, and that gap is goodwill. When the income result falls below the asset result, the market is telling you something blunt: these assets may be worth more in another use than inside this business. That is an uncomfortable signal, and many owners go a whole career without ever calculating it.

A practical note: the three approaches demand very different volumes of data, and the completeness of your data package itself moves the price. Every "let me check and get back to you" during due diligence is logged as risk, and accumulated risk converts into a discount or an earn-out clause. Preparing documentation is not administrative work. It is negotiation work.

Six factors that actually move the multiple

Direct answer: identical EBITDA can support very different multiples, and the difference nearly always sits in six places — customer concentration, owner dependency, bookkeeping transparency, growth, industry structure and transferability. These six are also the only levers you genuinely control before a sale.

1. Customer concentration. This is the first number buyers look at. If one customer exceeds 30% of revenue, or the top three exceed 60%, buyers treat it as structural risk, because nobody can guarantee those customers will not re-tender once the owner changes. The fix is not to fire your largest account. It is to grow the share of secondary customers year by year, convert verbal understandings into written agreements with terms, and make sure the relationship lives in the company — team, system, service process — rather than in the owner’s phone contacts.

2. Owner dependency. The next section covers this fully. The short version: buyers are not assessing how capable you are, they are assessing whether the company still runs when you are not there. In many Taiwanese SMEs those two things are in direct conflict — the more capable the owner, the more dependent the company, and the lower the multiple.

3. Bookkeeping transparency. Buyers can only buy what they can verify. Levers include normalising the books two to three years ahead, building real cost attribution so you know which customer and which order actually make money, cleanly separating related-party and personal expenses, and obtaining audited statements where appropriate. Slowest to take effect, largest in leverage.

4. Growth. Buyers pay for the future. A company with flat revenue for a decade will struggle to command a high multiple even with steady profit, while one showing three years of rising revenue and margin, with an explainable source of growth — new markets, new customers, new product lines, improved utilisation — prices very differently. The requirement is that growth be explainable and repeatable, because one-off pandemic or currency windfalls will be stripped out.

5. Industry structure. Your position in the industry cycle, barriers to entry, bargaining power up and down the chain, and regulatory or environmental trends all feed the multiple, and most of it is outside your control. What you do control is timing: negotiate while the cycle is favourable and you still have a growth story to tell, rather than after orders slide and you are simply tired.

6. Transferability. The most overlooked and most lethal. Do your critical assets travel with the deal? Is the plant owned or leased, how long is left, is it assignable? Is the key process documented or resident in one technician’s memory? Will ISO and customer-side certifications require re-audit after a change of control? Do supplier contracts carry change-of-control clauses? Every "not sure" becomes a price protection demand.

Put the six together and a pattern appears: none of them asks how much this company earns. All of them ask whether it will keep earning once this particular owner is gone. That is the core of valuation logic — buyers are not paying for past profit, they are paying for the certainty of future profit.

Why "the company collapses without the owner" cuts your valuation directly

Direct answer: buyers pay for future cash flow, and in a company that depends heavily on one person, the certainty of that cash flow collapses the moment that person leaves. Owner dependency is not an adjective. It shows up contractually as a discount, an earn-out, or a retention clause.

Start with Taiwanese data. PwC’s 2025 family business survey, covering 32 Taiwanese respondents, found that 94% of Taiwanese family members work inside the family business, while more than half of Taiwanese family businesses have no family governance mechanism at all — no family constitution, no equity entry and exit mechanism, no conflict resolution process. In the same survey, 47% named succession planning as a leading operational challenge and 69% named talent and leadership development. Those figures describe one condition: decisions and relationships concentrated in a few people, with institutionalisation lagging behind. In a good year that reads as efficiency. In a valuation it reads as a discount.

How do buyers test it? With plain but lethal questions. Who approves a quotation? Above what amount does the owner have to sign? Whose mobile number is stored in the procurement contact’s phone at your three largest accounts? Who builds the production schedule, and against what rules? Who makes the final call on a quality exception? Who negotiates supplier pricing? If more than half the answers are "the owner", then no matter how attractive the statements look, the buyer concludes he is acquiring an asset that requires the seller to keep showing up. Price reflects that, usually as a lower cash-at-close, a larger performance-linked instalment, and a two to three year retention plus non-compete.

Time to kill a widespread myth. Many owners treat "this place cannot run without me" as a moat worth points at the negotiating table. The opposite is true. In a buyer’s risk model, an irreplaceable individual is the highest grade of key person risk. The harder you prove that only you can decide, only you can calm that complaint, only you understand that old machine’s temperament, the higher the buyer’s discount rate and the lower the multiple. The moat has to belong to the company, not to you.

So how do you convert the owner’s capability into the company’s capability? One core move: turn tacit knowledge into explicit systems. This is far cheaper in 2026 than it was a decade ago. It works in three layers.

Layer one is making information and process explicit: quotation logic, cost structure, scheduling rules and quality criteria move out of the owner’s head into something queryable and handover-ready. We unpacked this in our piece on AI production scheduling and BOM management — the hardest thing to hand over in a traditional factory is rarely the equipment, it is the judgement embedded in scheduling, and once that judgement becomes rules and data it converts from a personal asset into a company asset. Our three traditional-factory AI automation case studies point the same way: what changes the underlying condition is not the tool, it is moving repeated judgement off a person.

Layer two is de-personalising customer acquisition. If orders depend on the owner’s network and dinner table, buyers will reasonably ask whether those relationships survive retirement. If some share of new customers arrives inbound through the website, through search, through content, that channel does not sit with any individual. It sits in the company’s digital assets. This is why "should we rebuild the website" is an asset question and not only a marketing question for a company in transition — an angle we covered in should an SME update its website in the AI era, with the build and content structure itself falling under our website and SEO service.

Layer three is making operations outsourceable. Whether a task can be handed to an external team against an SOP is the best possible test of whether it has been systematised. If day-to-day e-commerce and channel operations already run on documented process, dependency on any specific person is low. That is the hidden succession value of a service like e-commerce operations outsourcing: it forces the process onto paper.

Worth noting from the same PwC survey: 61% of family businesses globally see generative AI as a growth opportunity, against just 13% in Taiwan. In a valuation context that gap cuts two ways. Near term, it means most Taiwanese peers have not finished systematising, so whoever moves first holds a relative advantage. Longer term, buyers — financial buyers especially — will increasingly treat digital maturity as a proxy for transferability. That echoes the IBBA Market Pulse Q1 2026 finding that 67% of advisors see no material AI valuation impact yet while 12% already see upside: the market is forming a judgement, it has not concluded one.

A closing note on timing. Reducing owner dependency is a three-year project, not a three-month one. Systems have to survive a full peak and slow season, managers have to make and correct real mistakes, customers have to get used to speaking with a manager instead of the owner. Which is why valuation deserves attention three years before you want to sell, not the week a buyer knocks.

Next steps: who performs a formal valuation, and what to prepare

Direct answer: organise your own data first, then engage professionals for a formal opinion. In Taiwan the usual path is an accounting firm or financial advisor who knows your industry for valuation and tax structuring, a lawyer for equity and contract risk, and where formal documentation is required, a qualified valuer issuing a report under the valuation standards bulletins.

Who to approach. First, the transaction services or financial advisory arm of an accounting firm — the Big Four, including PwC Taiwan and KPMG Taiwan, plus the larger domestic firms, all field such teams, and their advantage is holding valuation, tax and deal structure in one place. Second, a credentialed valuation professional who issues reports under the valuation standards bulletins described earlier, which matters when the document must stand up in court, before tax authorities, in a shareholder dispute or at a board. Third, business brokers and M&A advisors, whose value is buyer access and execution, though remember their fee is usually tied to closing, so their valuation view carries a built-in position. Fourth, your long-standing audit accountant, who knows your books best and makes a good first stop — but if he is also the person who has been arranging those books, get a second opinion before you sell.

What to prepare. This checklist exists less to satisfy an advisor than to let you see your own gaps.

1. Five years of financial statements — balance sheet, income statement, cash flow statement, audited if possible, plus the latest trial balance.
2. An adjusted EBITDA schedule itemising excess owner and family compensation, personal expenses, one-off items, non-operating gains and losses, and related-party pricing differences, each with supporting evidence.
3. Customer and supplier analysis — three years of revenue and margin for the top ten customers, contract status and payment terms by customer, top ten suppliers and alternative sources.
4. Asset register — land and building title plus latest appraisal, equipment list with model, year, condition and original cost, lease agreements and remaining terms, inventory ageing.
5. People and organisation — org chart, tenure and scope of key managers, compensation structure, labour insurance and pension funding status, any labour disputes.
6. Compliance and risk — pending or potential litigation, environmental and safety records, certifications and permits with expiry dates, IP schedule, change-of-control clauses in material contracts.
7. A three-year operating plan — growth sources, capex plan, capacity utilisation, known order visibility, and the assumptions behind the forecast.

Assembling that package is itself a diagnostic. Most owners stall at items 2 and 6, because that is where the unaddressed history usually sits — and that is exactly where a buyer will press on price. See it three years early and you have three years to fix it. Let the buyer find it during due diligence and the only remaining variable is the discount.

Procedurally, a full valuation runs: define purpose and basis of value, gather and verify data, normalise the financials, select and apply approaches, cross-check and run sensitivities, issue the report. Steps two and three always consume the most time, and how long they take depends on how well your data is prepared. Which is the same point this article has been making throughout: the quality of a valuation is bounded by the quality of the data.

Disclaimer: this article is a general explanation of valuation methodology intended to help owners understand the mechanics and prepare. It is not a valuation opinion on any specific company, nor investment, tax or legal advice. Any actual business valuation must be performed by a qualified accountant, valuer or financial advisor who has reviewed real financial statements, contracts and asset records. All multiples and transaction statistics cited here are US or global public data and do not represent Taiwanese SME transaction levels; no public statistics exist for the transaction multiples of unlisted Taiwanese SMEs.

If the question you are really working on is the earlier one — how to strengthen the underlying business, how to build the systems, how to reduce owner dependency — that is where HappyCXO Studio can help. We do not perform business valuations. We help Taiwanese SMEs convert what only the owner knows into company systems and digital assets, so that on the day you want to hand over, you own a company that can be valued clearly. To talk through your situation, get in touch, or read the rest of the succession and business sale collection.