Many years ago, I learned an absolutely invaluable lesson in competition. I was employee #10 and working for a Boston based software company that developed the first Applicant Tracking System (ATS) that scanned resumes and created a digital searchable version of the resume. The company was Restrac.

This allowed recruiters and hiring managers to build searches against their open job requisitions to find the best candidates based on the skill keywords, educational background, job titles and experience in the digital resumes.

Our company co-founder was the guy who pioneered the concept of Contract Recruiting. He built out a contract recruiting agency across the US armed with contract recruiters with specialized recruiting experience (i.e., all the various subcategories in engineering such as electrical, mechanical, civil, software development, etc.).

They won multiple huge multi-year contract recruiting contracts with the big defense contractors who had won huge US defense contracts and needed to staff up with thousands of specialized engineers in a very short period of time. They were dealing with hundreds of thousands of resumes per year per contract customer.

Our founder thought that there has to be a better way than to manually code a resume for skill words, educational background and job experience in a homegrown database. Which was considered cutting edge at the time.

Being first to market does have its advantages in that we were growing fast over the first couple of years competing against some low end competitors. Then a new, formidable competitor from Silicon Valley arrived on the scene. They were based in Silicon Valley and were VC backed by Kleiner Perkins.

Their founder was a technologist that as an engineering hiring manager found the entire recruiting process to be highly manual and grossly inefficient. They developed an algorithm that automatically matched scanned digital resumes against open requisitions. The competitive company was Resumix.

They touted their “Patent Pending Artificial Intelligence”. Remember that this was back in the mid 1990s. AI was not even a thing back then. Their value prop and competitive positioning against us was speed. They could scan resumes faster than us and they could automatically match the best resumes for a job against an open requisition faster than us.

We would battle head to head with Resumix in all of the enterprise deals with F1000 companies that were receiving hundreds of thousands of resumes per year and filling thousands of highly technical jobs per year. We were really the only two options for these enterprise customers to consider and use.

Both companies were growing and prospering. I was learning how to sell enterprise class software deals. Then Resumix started to beat us. They were laying down the gauntlet and challenging us to head to head bake offs in a competitive customer’s headquarters. They did an excellent job of influencing the decision criteria (e.g., MEDDIC) and convincing the customer that speed and time to fill were the most important factors that they needed to consider.

For the first time, we were losing to Resumix more than we were beating them in the largest enterprise competitive deals. It was painful and I’ve always been a really sore loser. We needed to find a way to turn the tables and seed/influence the enterprise deal decision criteria to our solution’s unique differentiator. I learned how to crack the code successfully through a brutally competitive sales cycle at Intel against Resumix.

This was your classic enterprise sales cycle. It was over a year in length. There was an RFP that was ridiculously onerous and long. Intel is an engineering centric organization. Suffice to say that this was an excruciatingly detail oriented, data driven evaluation and decision process. It all culminated in a two (2) week head to head bake off in Intel’s corporate headquarters in Santa Clara, CA.

The rules and scoring rubric document given to both vendors were quite explicit and long. Intel wanted this bake off evaluation to be 100% objective and provide a level playing field. Neither vendor/sales team was allowed to touch the actual computer keyboard that the Intel HR employees were using during the bake off to simulate how they would do their jobs using both of our software systems.

Intel would scan the same resumes into both software systems. They the Intel recruiters would perform searches and matches against their most difficult to fill engineering job requisitions. Then Intel would compare the search/matching results against the short list of candidates that they interview for those actual jobs and the people that they had actually hired for those jobs.

Needless to say that we decisively won the Bake Off against Resumix. What we learned is that search accuracy and quality of search results were way more important than speed. And our search engine did a much better job of finding new and esoteric skills that they wanted to hire against for their next generation semiconductor engineering jobs.

This exposed the fatal flaw in Resumix’s proprietary algorithm and so called competitive advantage against us. We indexed and were able to search the entire text of the scanned digital resume. Whereas Resumix had a black box skill lexicon table that they would use to tag and match resumes against open job requisitions.

What we learned is that the Resumix approach was fundamentally flawed in that if the skill words on a resume at the time they were being scanned into the system did not exist in their black box lexicon, there was no way to retrieve or match them because in the interest of speed they only searched/matched against a small subset of the entire resume.

We now had a competitive blueprint for success against Resumix. We became experts at seeding/influencing the decision criteria on all of the enterprise deals over the next few years. We would ask questions such: as “Do you ever anticipate having a need to search for skill words on the resume that were not considered important at the time of being scanned?” “How important is having the ability to search for any keyword on a resume at any time and being able to find that candidate?” “Do you know at the point of scanning in resumes all of the keywords that you might use to search/retrieve that resume at a later date?” “How would that impact your confidence that you are getting the best, most accurate resume search results?” “Is search accuracy more important than the how fast you get search results?

This culminated in us crushing Resumix in head to head Bake Offs and ultimately allowed us to successfully IPO.

Good selling!