My first employee

I just hired my first employee, a Ph.D. Data Scientist. Here's my long-winded story about why this made sense for my Marketing Agency before I even knew it did.

In September 2015 I met a friend of a friend for happy hour. He owned a successful agency in Seattle and I was hoping to get some advice for running my own. I was in the two month window between deciding to start my own marketing agency and getting my first client.

After proclaiming I was starting an agency, he asked about my focus area. At the time I thought I wanted to specialize in agile user research and take a chunk out of the lucrative traditional research market in Seattle. I had a lot of experience in this space and I had done a couple research projects over the summer, testing my chops outside of Microsoft.  

The research I was envisioning was fast, agile, startup-style — and it was mostly qualitative.

So when he asked the next question, we were both shocked by my answer.

He asked, “Who are the first 5 employees you’d hire?"

I said, “I don’t know about 5, but I know my first, I need a Data Analyst.”

The silence that followed felt super awkward. I thought, perhaps it’s the beer because that answer makes no sense. I don’t need a Data Analyst for the type of research I was just talking about.  

Yet, something about having a Data Analyst was a need-to-have for me and my work history explains why.

Over my career at Microsoft I was always drawn to data, I needed it to help make decisions and when I couldn’t access it, I invented ways to get it.

In 2008 after joining the Bing Engineering team -- making the leap from MSN Marketing, I quickly realized I couldn’t access any Bing data. It was stored in Cosmos, a system only accessible by Bing Engineers who knew how to query the database with the proprietary coding language called SCOPE. 

I was at a disadvantage to my pedigreed Computer Science peers who could bring data to the table, so I championed what's now known as the 'Search Data portal' to be built on top of Cosmos -- removing the barrier that was SCOPE. Bing Marketers loved me, my plan would help them out. 

It also made sense for executives, they'd be able to access the data as well. It was an easy sell and I secured the budget. This was my first real project as a Bing PM but I quickly learned I was an outsider and was not liked by the Cosmos team because they “owned” the data. One day I showed up to work to find a 2 inch thick C++ coding book on my desk, SCOPE is based on C++. Someone was giving me a not so subtle message to learn their ways. I opened the book and my eyes glossed over -- but instead of ignoring their hint I decided to give it a shot and signed up for C++/SCOPE training.

For those of you who don’t know me, I went to Pharmacy school, I can run circles around engineers explaining the kidney's role in dehydration, but coding was my Kryptonite. In the SCOPE class I wasn’t able to follow along with simple directions — everyone was speaking a foreign language. The instructor had to slow down the class and look over my shoulder the whole time. After 20 minutes, I said, “I feel like I signed up for the wrong college course,” and I excused myself.

I decided to continue down the path of creating the Search Data portal. The Cosmos team declined to collaborate. That sucked. But even without their help I was still making progress because I had hired an outside company to build it. After the Cosmos team realized what was happening, they tried to race me to the finish line by building a portal themselves. Rewriting my spec (my detailed plan) and presenting it to the VP as their own. At which time he realized there were two different efforts. Thank goodness. In the end, my solution was cheaper and faster to deploy and the Cosmos team's effort was squashed.

The Search Data portal launched and information about our customers was now more accessible to everyone at Bing.

I had my data and all was right with the world. Until I came across another roadblock the next year.

In 2009 I was a UX Program Manager -- UX PM’s were in charge of designing new features on Bing.com. I would come up with ideas, get a designer to bring them to life and review the ideas with my developer. Many times the developer would declare the ideas "bad", and they wouldn’t build them. They also had ideas I didn’t think were good and I wouldn’t plan them so it went both ways. Deciding what features were built usually came down to who was stronger willed.

I thought, there has to be a way to bring more data into the planning process. I wanted to get designs in front of customers and collect feedback before meeting with my developer. But the planning process at Bing was accelerated. We had just a few weeks to go from ideas to specs, so carving out time to include research would be challenging. Results needed to be available within hours, not days or weeks. 

I set out to develop a solution and I created a new platform called “Quick Turn Research”. For the first time, it allowed PM's to walk into their first design review with data to back up decisions. 

After years of maturity the platform is now called UX labs which Brent Schnabel blogs about. 

One of the last things I did before leaving Microsoft in 2014 was demo UX Labs at the Microsoft Exec Retreat to 150 of the top executives. A proud accomplishment.

I again had my data. (Smile).

This next story is not about me needing data, it's about my VP needing it for a Bill Gates review and I rose to the challenge. 

In 2010, the VP of Bing had a Bill Gates strategy review looming and was stressed because he didn't have the data he wanted to anchor it. He wanted to have a clear breakdown of all searches coming through Bing.  The main issue is that many searches are ambiguous — 'Blue moon' could be the beer, the restaurant, or the color of the moon. So it could be counted three times; over inflating the totals. 

The VP issued a request to his employees -- is there a PM who can get more accurate data in the next two weeks? 

My manager asked what I thought. "Oh yeah, I'm in."

Because I'm not trained in CS, I don't develop software to solve problems, I leverage existing software -- perhaps that's what allows me to approach problems differently. Or it could be my inventor DNA. My grandpa invented the first milking parlor after he broke his back falling out of a grain silo and couldn't milk cows by hand -- he chased people off his property with a shotgun who were trying to steal his design.  But he never profited from it. That's ok, he's 92, happy, and still farming, spending more time with his bees these days than with the cows. But I digress... 

In this case my invention was not physical. I invented a new research method — using people to label searches one-by-one. This meant people could label 'Blue Moon' as food, restaurant, or astronomy but not all three. It was a good idea in theory but we didn't know if it would work so we had to act fast. Together with the VP, we quickly developed a list of 144 categories that each search could be labeled (e.g. Restaurant, Hotel, Celebrity, etc…). Then I took a random sample of Bing searches and had multiple people label each one using the Amazon Mechanical Turk platform. It wasn't an exact science but I was able to prove the method passed the sniff test. It was good enough for a successful strategy review with Bill Gates. It was also the first time the VP had a clearer view of the Bing searches by category and my work earned me a Gold Star award, a one time bonus worth thousands of dollars.

Following this pattern of needing and creating access to data throughout my history at Microsoft, I realize now that my gut knew that having a Data Analyst on my team was a missing piece even before my head had figured it out -- it just took one beer for my gut to talk.

Three months ago I had enough client work and money to sign a 3 month temp-to-hire contract and I found a Data Analyst (or rather he found me), Yev Marusenko. A Ph.D. Data Scientist no less, who also has a marketing background and can navigate the Internet like nobody's business. Our three month trial period is coming to an end and I’m pleased to announce he has accepted the offer to join me at OnTheGrid full time. 

Yev is providing a level of precision to SEO and SEM that I didn’t know was possible. We're able to measure just about everything. I coined the term for our latest campaign, “Hyper-targeted advertising”. I don’t know anybody else who’s working at this level of precision so I felt a new term was justified. The other day Yev and I were geeking out over data as our new campaign was starting to come to life. Some poor 25 year old woman (who we know), was shown our ad on Facebook 62 times! We don’t exactly know why she saw anything on Facebook that many times, and I know, it sounds really creepy, but how cool is it that we can advertise to one specific person?! (Disclaimer, I didn’t create these ad platforms, we’re just using them at an advanced level which the platforms likely weren’t designed for. A loophole? How far can we push it? I'll keep you posted.)   

Having a Data Analyst and this level of insight is exactly what I needed before I even realized it.

Everything is as it should be.

#WomenInTech #Startups #Leadership

Jessica Jobes is Founder of OnTheGrid a Seattle Marketing Agency.