I'm currently in the process of making some changes to the RJMetrics voice used in marketing communications. I feel like we're too conservative in blog posts and white papers and don't make effective use of humor, which in my experience is one of the best tools for content marketing.
This exercise has me thinking a lot about corporate voice. How does a startup decide how to present itself on its marketing site, blog, 1:1 emails, in conferences, etc.?
There is a lot that goes into answering this question, and most of it I don't want to touch on here. The often-overlooked element of voice, though, is this: your startup's voice should reflect the essential culture of your company. (If you haven't read Ben Horowitz's post on company culture, there is none better. Read it.) This is a hard thing to do for the same reason that getting your culture right is a hard thing to do—it's just not that damn obvious.
Which is why I give a huge amount of credit to Bob and Jake at RJMetrics. While I think they have erred on the conservative side of voice for marketing communications, they got the culture part right. Take a look at this gem found hidden within our Voice document on Google Docs:
We apply our focus on data to other areas of our life. We analyze the lyrics of our favorite songs, understand the business models of our online dating sites, and try to optimize our commutes. Whether we are talking about RJMetrics, data in general, or our personal lives, we use the same voice because this is the way we actually think.
I hope they don't mind that I'm pasting internal company docs onto my blog; I just thought this was so great. We speak like data geeks because we are data geeks.
Chris Dixon has a way of stating things that the industry widely believes to be true, but doing it in such a simple, straightforward way that it still forces you to stop and think. I loved his recent post on the frontiers of tech:
Today, the tech hobbies with momentum include: math-based currencies like Bitcoin, new software development tools like NoSQL databases, the internet of things, 3D printing, touch-free human/computer interfaces, and “artisanal” hardware like the kind you find on Kickstarter.
It’s a good bet these present-day hobbies will seed future industries. What the smartest people do on the weekends is what everyone else will do during the week in ten years.
I would add things like: the spread of developer tools to normals turning more professionals into hackers (git for legal docs, etc); analysis layers on top of big data (personal data mining as well as business); and ecommerce-as-a-service (tools to make selling things online so easy that a huge % of the population become small business owners).
All of these things are incredibly personally exciting to me. You know the stereotypical "I thought we'd all be wearing jetpacks by now!" 1950's vision of the future? I think this stuff is better. Every one of these technologies is egalitarian (or enables egalitarian principles). Which makes me think back to Future Perfect and its vision of a peer progressive society.
I'm so happy I get to go to work and be a tiny spec of this story.
Good way to start the week.
Mark Suster's blog post today really hit me. It's a topic I've thought a lot about in the past several years—how do you manage to be an entrepreneur and still maintain healthy personal relationships?—and one that I know a bit about.
I, like Brad Feld, was married in my early twenties. Also like Brad, my marriage didn't last. At the time I was a consultant, and during the three years that my ex and I were married, I was out of town every single week for 4-5 days. In retrospect, I can't believe that 25-year-old me thought that was anything other than a poor life choice.
But the thing that I really loved about Mark's post is actually a quote from the book. In it, a woman describes how her life changed after a harrowing accident left several members of her family severely injured:
I appreciate kindness more. I recognize kindness more often. I do stop to smell the roses. I give people a break. I give people the benfit of the doubt.
I ask more questions. I forgive more. I am more open to ideas. I have more faith in mankind. I am more tolerant. I make an effort to understand a situation or what someone is trying to tell me.
I am more more sensitive to others who are in pain. I want less responsibility. I want less materialistically. I enjoy purging and freeing our lives of stuff.
I loved this quote, and I completely identified with it. Getting divorced was the first big failure of my life. Not that everything up to that point had been an incredible triumph, but I had never really screwed the pooch before. Failing big, and failing publicly, taught me a degree of humility that I had just never possessed before.
The changes in my personality and outlook since then have been very similar to the ones described in the quote above. I'm more patient, invest more in people, and put much more effort into listening. I see people. I temper myself more, and am more self aware.
We learn through adversity all the time in our professional lives; it's wonderful to hear someone talk about it in the context of their personal life. And I think it's really interesting that whether you're talking about divorce or close family injury, the thought processes are often the same: slow down, be kind, simplify, and focus on what matters.
I've been a data analyst for 15 years. I am, by now, more than a data analyst, but it is still the heart of what I'm about.
Coincidentally, that puts the start of my analyst career right about at the launch date of one of the most significant analytical tools since the abacus: Excel 97. Excel 97 was an amazing product in its day. If you go back in time, pivot tables were just a shade below pure magic. And, with the introduction of Access as a part of Office, Normals actually had a desktop-based SQL engine.
And VBA! Oh my god. I learned to program in VBA, made functional in Excel 97. Business users could, for the first time, automate the things that they previously had to spend their own time doing. Considering that our economy is founded on increasing employee productivity, giving Normal knowledge workers the ability to automate their own tasks is pretty darn significant.
The flow would go something like this: collect data in Access >> export to Excel >> run pivot tables and more >> make charts >> paste into PowerPoint >> print >> impress boss. In 1997, this was groundbreaking stuff. And the fact that it all worked together (reasonably successfully if not pleasantly) was a triumph of product integration.
Office 97 unlocked the possibility, for the first time ever, of doing legitimate data analysis at price points that your boss would actually pay for. And price is a critical factor in innovation—I don't care that SPSS has always been more sophisticated than Excel, as almost no one actually has access to it (and even if they did, it's well beyond the capability of Normals).
But Excel hasn't really changed since 97. VBA has been extended to account for the "Internet". Row limits have been increased. But otherwise, every meaningful user-facing feature has been UI-focused. Take a look at the Excel Wikipedia page and go through the version updates: nothing significant.
The reason for this is pretty straightforward. In 1997, Excel hit the end of where it was going to go as a technology and a product. It provided all of the tools a business analyst needed to do the things that a business analyst needed to do.
But as Excel, my old friend, has aged. It is has developed smile lines. Crow's feet. While it solves its problem—discrete business data analysis—very well, it does not do a good job with the larger problem: continuous business data analysis. Excel is good at answering a single question at a single point in time, but it completely, totally fails when it comes to ongoing process and decision making.
- Every Excel worksheet is a standalone entity. It doesn't have any relationship to the core data systems that run an organization: accounting, sales, point-of-sale, etc. So every time you export data to be analyzed in Excel, it is outdated by the time you open up the file. And you have to start over (a.k.a. "Save As...") in a month when you have to run the report again.
- There is no single point of truth. I was a data analyst at General Electric for a summer, and the experience was terrible. You want some type of data? Ask 8 different people to find out who has "the spreadsheet" and have that person email it to you. Repeat for every additional type of data you need. If you ever want updates, email all of those people again. If you actually want the data from disparate sources to relate to each other, you had better have a god to pray to. Additionally, any given metric could be defined a number of different ways. Is my "revenue" the same thing as your "revenue"? Often not. This makes it impossible to tell the derivation of any given metric without traversing through an entire spreadsheet.
- There is no standardized "way to construct a spreadsheet". So trying to read someone else's workbook is basically impossible. This creates incredible amounts of job security for sub-par analysts who never get fired because if they did, no one could ever take over their spreadsheets.
I need to stop the list there or I'll never finish writing this post and you'll never finish reading it. Essentially, Excel was too powerful. Users, desperate for solutions, used it as a tool for operational—continuous—problems as opposed to discrete decision analyses. And it failed utterly, costing businesses billions and billions of dollars in lost employee productivity and lost insight. At Deloitte, where I was a technology implementation consultant for five years, our job was to replace overgrown spreadsheets. Businesses would use spreadsheets to operationalize process, and when the process ground to a halt they would call us in to custom develop software to replace the functionality. We were paid a lot of money to do this.
Fifteen years later, we finally have a real attempt to solve this problem. Not some customized Cognos installation that has a "web front end": a real, productized, way of solving operational data analysis. GoodData, Birst, Bime, and RJMetrics are multi-tenant (cloud/SaaS) solutions that all have a single goal in mind: kill operational spreadsheets.
Each of these tools provides extract-transform-load capabilities to get all your data into a single repository and keep it current without having to email 8 people every month. They all provide a single point of truth, including a complete explanation for how that truth is derived. And, they solve the "bad analyst" problem in a different, somewhat unexpected, way. If there is an innate separation of data and analysis (which is something that all data warehouses force on you), it's very easy to rip up any given analysis and start from scratch with the raw data. Beautiful.
On February 4th, I'm joining RJMetrics as their Director of Marketing. I was an early customer while at Squarespace in 2010 and have been a champion ever since. I'm excited about participating in this industry in its very early stages, working with an incredibly talented team, and building groundbreaking software.
I wish I had access to tools like this when I was 17 and first learning how to use VLOOKUP(). I want to build them for all of the 17-year-olds out there now.