Exhibit G · Revealed strategy
A strategy deck is a claim. An open requisition is a budget somebody already signed. This reads any public job board and returns the investment mix by region and function, so you can set it against where the revenue actually comes from.
Companies tell you where they are going in the language of ambition. They tell you where they are actually going in the language of open requisitions, because every one of those is a headcount budget that somebody already approved and is already paying for.
Job boards are public. So the mix of those requisitions, cut by region and by function, is a read on capital allocation that nobody has to disclose to you and nobody bothers to obscure.
Elastic reports that customers outside the United States are 46% of revenue. Its live board is 63% non-US roles, seventeen points ahead of the revenue share. On its own that reads like international investment running ahead of international revenue, which is what you would want to see.
The function mix says something more specific. Outside the US, more than half of open roles are engineering or product. Inside the US, nearly half are field and go-to-market, and the US alone carries more field requisitions than the whole of EMEA. So a meaningful share of what reads as international expansion is engineering capacity placed in lower-cost hubs rather than revenue capacity placed in market.
That is not a criticism. Placing engineering in Barcelona, Lisbon, Athens and Bangalore is a deliberate and defensible strategy. It just means the headline number and the underlying commitment are answering different questions, and only one of them tells you where field capacity goes next.
Every location and title is classified by an ordered list of regular expressions rather than an LLM call. Two reasons. A finding has to survive being challenged line by line, and a taxonomy that drifts between runs makes any time series meaningless. The cost is that the rules need maintaining; the benefit is that the mapping is auditable and identical every run.
A sales engineer is field capacity, not research and development, so the field pattern is tested before the engineering one. Revenue operations sits in field; revenue accounting sits in general and administrative. Those are judgment calls that move the numbers, so they live in one readable file rather than inside a prompt.
Anything unmatched lands in an explicit unmapped bucket and is printed with the results. The first Elastic run left sixteen percent of US titles unplaced, which quietly flattered the field share until the taxonomy was extended to cover account-based marketing, regional vice presidents, alliances, FP and A, and one requisition written in Japanese. Silent bucketing is how an analysis ends up confidently wrong.