August 10 in business history: Aspirin, Starbucks, FriendFeed, and Alphabet show what a bet needs next (2026)

On August 10, 1897, a Bayer chemist produced acetylsalicylic acid in a chemically pure, stable form. On August 10, 1987, Howard Schultz documented the handoff that put Starbucks’s name, roasting plant, and six Seattle stores under his control. On August 10, 2009, Facebook agreed to buy FriendFeed, including a team whose product ideas it had already begun to copy. On August 10, 2015, Larry Page split Google from the rest of the bets into a new parent company called Alphabet.

Those dates share a less obvious subject than corporate history. Each decision moved a promising thing across a boundary: lab result to product, asset to operating model, team to platform, portfolio to governance. The first proof made the next obligation unavoidable. That is the useful mirror for today’s launch, acquisition, or reorganization: identify the handoff before you celebrate the signal.

1. Bayer: a discovery is not yet a product

August 10, 1897 — product-development milestone.

Felix Hoffmann, a chemist in Bayer’s laboratory, synthesized acetylsalicylic acid by acetylating salicylic acid with acetic anhydride. The laboratory record gives the date, and the result mattered because the compound was chemically pure and stable. Bayer’s history of Hoffmann records the synthesis and the later development work; the German Patent and Trade Mark Office’s Aspirin history reproduces the August 10 laboratory-log evidence.

The chemistry still had to survive contact with a business. Bayer says large-scale studies confirmed the substance’s efficacy and tolerability, then describes the work required to develop a cost-effective production process. In 1899, the company launched it under the trade name Aspirin, initially as a powder in glass bottles. Bayer’s account makes the outcome unusually clear: the product made the Bayer name world-famous, and Hoffmann’s discovery was already a worldwide success by the time he retired in 1928.

The point is not that a good invention needs years of patience. It is that the invention and the product have different proof requirements. The August 10 result established a stable compound. It did not establish a manufacturing cost, a clinical case, a brand, or a route to market. Those were separate decisions, each capable of killing the original promise.

The mirror for today: when a prototype, model, or technical demo works, name the next proof in operational terms. Is it yield at target cost? Reliability under load? A repeat purchase? A regulatory path? If the answer is only “more users” or “more attention,” the team may still be measuring the lab result rather than the business.

2. Starbucks: the acquired asset needed a new operating model

August 10, 1987 — acquisition handoff.

Howard Schultz had left Starbucks in 1985 after the founders declined to shift the business toward the Italian espresso-bar experience he had seen in Milan. He started Il Giornale, opened three locations with coffee made from Starbucks beans, and then got a chance to buy the company he had left. In March 1987, Starbucks’s founders put its name, roasting plant, and six Seattle stores up for sale. By August, Schultz had raised the $3.8 million needed for the deal. Starbucks’s official history documents that sequence.

The August 10 artifact is a memorandum from Schultz to Starbucks employees about Il Giornale’s acquisition. The same historical page identifies the transition detail: existing staff received offers under the new ownership, with the same pay and benefits effective August 17. That detail matters more than the purchase price. A transaction changes the business only when customers, employees, supply, and brand can cross with it.

The buyer also made a revealing choice about identity. Just two years after Il Giornale began, it acquired Starbucks’s assets and adopted the Starbucks name, continuing as Starbucks Coffee Company. The acquiring concept supplied the espresso-bar model; the acquired company supplied a name, roasting capability, stores, and a relationship with existing staff and customers. Schultz was not merely adding locations. He was choosing which parts of each company should become the operating system.

This is why the deal is a better mirror for an acquisition than the usual “strategic fit” sentence. The fit was not self-executing. It depended on deciding what to preserve, what to change, and what employees would experience on the first day. A buyer that treats the target as a bag of assets may discover that the value lived in the habits around those assets.

The mirror for today: before signing, write two lists. The first names what the target owns. The second names what customers and employees actually rely on. If the lists differ, the integration plan is not an appendix to the deal; it is the deal.

3. FriendFeed: an acquihire can buy ideas before it buys revenue

August 10, 2009 — Facebook agrees to acquire FriendFeed.

FriendFeed was a real-time feed service built by former Google employees including Paul Buchheit, who had created Gmail, and Bret Taylor, who had helped launch Google Maps. It gathered updates from social networks, blogs, bookmarking services, and RSS feeds. By the time Facebook bought it, Facebook had already borrowed features FriendFeed had popularized, including the Like button and a stronger emphasis on real-time updates. TechCrunch’s contemporaneous acquisition report records both the announcement and the product overlap.

Facebook’s quoted press release said that all FriendFeed employees would join Facebook and that its four founders would take senior roles on Facebook’s engineering and product teams. The release did not disclose financial terms. A contemporaneous TechCrunch report on the consideration later put the figure at $15 million in cash plus roughly $32.5 million in Facebook stock, while noting that the stock valuation depended on an employee-share transaction and vesting over several years. The distinction between disclosed and reported terms is worth keeping: a deal can be strategically legible before its price is public.

The acquisition therefore had three possible assets: the live product, the product ideas, and the people who could make those ideas travel inside Facebook. Facebook had already shown that it could reproduce individual features. The harder question was whether it could retain the team’s speed and judgment after moving them into a much larger system.

The outcome supplies a useful check against a flattering acquisition story. In March 2015, FriendFeed’s own team wrote that usage had steadily declined and that the service would shut down on April 9 after roughly five years under Facebook. The archived shutdown announcement is direct about the reason: the community had become a fraction of what it once was. The product did not become a durable standalone network. That does not prove the acquisition failed; it shows that the target’s value had to be judged somewhere other than the target’s continued existence.

The mirror for today: decide whether an acquisition is buying revenue, distribution, a team, a capability, or a learning advantage. Give each asset its own success measure. If the product may be retired, specify how the buyer will prove that the people or ideas transferred value before the shutdown becomes the only visible result.

4. Alphabet: a reorganization is real only when accountability moves

August 10, 2015 — Google creates Alphabet.

Larry Page announced a new parent company, Alphabet, with himself as CEO, Sergey Brin as president, and Sundar Pichai as CEO of Google. Google would become a wholly owned subsidiary; its shares would convert into the same number of Alphabet shares with the same rights, and the two share classes would continue trading on Nasdaq as GOOGL and GOOG. Page’s original announcement, “G is for Google,” said the structure would separate Google’s core internet products from businesses that were “pretty far afield,” while giving each company more independence, focus, and accountability.

The decision addressed a management problem created by success. Google had search, advertising, Android, and Chrome alongside projects in life sciences, longevity, moonshot research, and venture investing. A single operating label could make a portfolio look coherent to outsiders while leaving capital allocation and executive responsibility blurry inside. Alphabet was an attempt to make the boundaries explicit and to give Pichai a narrower job at Google while Page and Brin managed the wider portfolio.

The test is visible in current reporting. Alphabet’s 2025 Form 10-K describes the company as a collection of businesses, with Google reported through Google Services and Google Cloud, while non-Google businesses are grouped as Other Bets. It says Other Bets operate as independent companies, some with their own boards and outside investors, and range from X’s research-and-development work to Waymo’s commercialization. The SEC-filed 2025 annual report supplies that description.

The numbers show why the separation matters. In Q4 2025, Google Services reported $95.862 billion in revenue and $40.132 billion in operating income. Other Bets reported $370 million in revenue and a $3.617 billion operating loss. Alphabet’s SEC-filed Q4 2025 results make the portfolio’s different clocks visible: a mature cash engine can fund bets whose economics are still being tested, but the bets do not disappear inside the core business’s margin.

The mirror for today: a reorganization earns its keep only if it changes who decides, who reports results, and which losses can be stopped. If the new boxes leave capital, talent, and escalation paths unchanged, the company has edited its chart rather than its operating model.

The manager’s test

Four August 10 decisions leave four questions for a live decision:

What exactly has been proven? A stable compound, a target’s assets, a specialist team, or a portfolio that has become too broad for one operating unit?
What handoff follows? Manufacturing, employee transition, product integration, or governance?
Where will the next failure show first? Unit cost, customer behavior, talent retention, or segment economics?
What should remain separate? Quality from discovery, the brand from the retail format, the team from the legacy product, or the core business from experimental bets?

The August 10 pattern is not a warning against bold moves. It is a warning against counting the first proof twice. A demo does not pay for manufacturing. An acquisition announcement does not integrate people. A talented team does not guarantee a surviving product. A new parent company does not create accountability by name alone.

Before approving today’s launch, acquisition, or reorganization, write the next obligation beside the first signal. Then attach one number or observable behavior that would tell you the obligation is being met. That is the part of the decision history tends to hide—and the part managers have to make visible.