Traditional DCF and comparable-transaction multiples routinely fall apart in front of tech companies. Too much value sits outside the financial statements, and the value that does sit inside is often mis-measured by accounting convention. Here are three blind spots we've seen again and again in robotics, AI, and SaaS advisory work.
1. The capitalized-R&D trap
Under K-IFRS (and equivalents in most jurisdictions), development costs can be capitalized when specific criteria are met. In practice, this capitalization often destroys, rather than protects, valuation.
Two reasons. First, capitalized R&D returns to the P&L as amortization. The buyer's forward EBITDA projections honestly reflect that amortization, and the EBITDA-multiple valuation drops as a result. Second, at Purchase Price Allocation, the buyer remeasures capitalized R&D. Any gap between book value and remeasured value becomes a seller-credibility issue.
Early in every engagement we run a "consistency check between R&D accounting policy and technical substance": does the capitalization policy actually match what was built, does the amortization schedule align with the technology's economic life. When it doesn't, we discuss reversing the capitalization before diligence begins. Sometimes you unlock valuation by uncapitalizing.
2. Quantifying talent density and attrition risk
Most of a tech company's assets sit inside people. Those assets never make it onto the balance sheet, and in valuation discussions they are usually handled with the qualitative sentence "key-person retention is critical." When the buyer fails to quantify this risk, sudden post-close attrition can collapse the deal's economics.
We attach three standard documents to the diligence package.
- Key Person Matrix — contribution, replaceability, and market re-hire cost (by tech stack) for the top 10–20% of headcount.
- Retention Cliff Chart — contract expiry dates, RSU/option vesting schedules, and 12/24/36-month attrition scenarios post-close.
- Rebuild Cost Model — time and cost to rebuild an entire team if it walks. This becomes the basis for earn-out structuring.
With these three documents on the table, buyers reach for earn-outs and retention packages rather than valuation discounts. That is a materially better negotiation for the seller.
3. Quantitative evidence for data moats and network effects
"We have a data moat" is common in pitch decks and rare in diligence deliverables. Without quantitative backing, moat claims do not translate into multiple premiums. Three metrics carry the weight.
First, data stock-to-flow ratio. Total accumulated data divided by data ingested in the last six months. A lower ratio (relatively large stock, steady flow) evidences a durable moat.
Second, data reproduction cost. Time and money for a competitor to build a dataset of comparable size and quality. Labeling unit cost, domain access rights, regulatory approval timelines belong here.
Third, the slope of the model performance–to–data curve. How many percentage points does model performance gain when data volume doubles? A shallow slope means the moat is theoretical. A steep slope means it is real.
These metrics are validated by the buyer's technical DD team, not their finance team. The moment a boutique advisor prepares these documents for the CTO's audience, the deal's center of gravity shifts from financial to technical diligence. And that is when tech companies finally receive the multiple they deserve.
Closing
DCF and comps are tools. When the tools cannot see the assets that actually create value, the advisor's job is not to replace the tools but to translate the assets into a form the tools can read. We have systemized that translation into three standard documents.