I went Back to the Future and I’m here to report what I found.
It was Thursday, April 18, 2013, from 9:00 until 4:00 at the San Francisco Marriott Marquis Hotel in conjunction with the eMetrics Summit. It was the eMetrics Executive Roundtable and I learned three very important things.
Seven years after launch, it was time for another foray into the unknown so I hosted another gathering of the initiated, resulting in 101 Things You Should Know; A Report from the eMetrics Analysis Symposium, Summer 2009.
But by 2013, things had changed so much, so fast, it was time to regroup and re-gather a smaller, more established, more accomplished set of individuals to take a peek over the horizon. The eMetrics Executive Roundtable included more than a dozen senior analytics types with real-world, hands-on, enterprise level customer experience experience. The topics on the table were: Where are we? What’s the hard part? What’s next?
The conversation ran from people to process to technology and the areas of concern were all predicted back in 2002:
- We have enough technology, we just aren’t sure what to do with it.
- Wrangling new data types is the new normal so it’s time to give up on technical perfection and focus on insight development and management.
- Corporate culture is crucial and change management is required. “Teaching people to fish” – to analyze data for themselves – is not as easy as hoped.
- We will always want to know which half of our advertising dollar is being wasted.
- The Internet democratizes brands. You stand for what the public says you stand for
- Finding and retaining the right people is a challenge.
The conversation brought forth many great snippets of advice for dealing with each of the above, but they were tactical and will be the subject of numerous blog posts to come. While valuable, none were as ground breaking as the three I things I actually learned this time.
1. You No Longer Need to Know the Question Before You Start Collecting Data
This resonated so deeply that it is still ringing in my ears. I was compelled to offer it up as a new definition for Big Data. The three V’s are nice (Volume, Variety, Velocity) but everybody gets hung up on volume: How much do you need before you can call it Big? How much data does it take to change a light bulb?
The amount just doesn’t matter. Whether it’s Stéphane Hamel’s, “That which no longer fits into Excel” or it’s measured in petabytes, yottabytes or brontobytes, the quantity is immaterial. What matters is that we can afford to capture and mange more than we could before.
Classic database design says the first thing you do is, “Determine the purpose of your database.” Then you can go out and organize the data that will answer your questions. And then divide the information into tables and columns with primary keys to uniquely identify each row, set up the table relationships, refine your design and apply normalization rules.
If you don’t know the questions you want answered, you cannot get off the ground.
Why? Because in a world of limited resources (disk space and processing power), careful planning was central to allocating those resources.
When Hector was a pup, I sold business computers. Our regional professional organization started selling Cado computers, warmed up to Data General and finally graduated to selling IBM System 34‘s and 36‘s.
One day, the sales team was driven down to the local Prime Computer offices where we got a demo of the new machine we’d soon be selling: the Prime 850. This 32-bit bad boy was a multi-stream, parallel processing wonder with ultra-high density MOS memory packing 64k bits on a single chip and could support 128 terminals – at the same time. It was a smokin’ hot machine.
After the chalk-talk about what parallel piping meant to number crunching, we actually got our hands on the actual device. Well, OK, we got to look on as our host sat in front of an amber screen, paused at the command prompt said, “Watch this. I’m going to ask for the phone numbers of all our customers who have purchased 550’s, 650’s or 750’s in the Southwest whose maintenance contracts are about to expire. It’s as simple and as fast as this…!”
We all murmured our admiration that he would try such a query on a multi-terminal machine. Such confidence. Such raw power!
And then we waited.
After five minutes – which is an interminably long time when you don’t know whether you should still be impressed or not – he got up and left the room with a cloud over his head.
After another five minutes he returned and explained that a new secretary had been asked to look up a phone number … and she did … on her terminal … at the command prompt … inarticulately querying seven databases at the same time and bringing the machine to its knees.
The demo was not nearly as impressive after that.
Today, with dirt-cheap storage and distributed computing on clusters of computers, the resources are not nearly so limited and the ability to pre-process unstructured data into structured data – fit for online analytical processing – is at hand.
As a result, potentially interesting data can be collected at will until such time as somebody comes up with the question du jour. Data collection is now a matter of course and we can really focus our attention of asking better questions more often.
2. The Website is No Longer an Oddity
In the beginning, the corporate website was a skunk-works project that a couple of guys in IT played with. Later, marketing discovered it and a small group of marketing and technical people kept it going because it was so cool.
When it started costing real money, management tried to find the right home for it. Large retailers did their best to ignore the whole thing until stores started to complain that these nerds were sucking up resources and cannibalizing sales. New silos of customer experience and behavioral data were born along with new political rivalries.
In 1995 I was brought in to explain the World Wide Web to a group or marketing executives at Sears. I told them that they would be soon selling everything online. There was outright laughter in the room. In an avuncular fashion, my host gently informed me that the current CEO was doing just fine even after killing the iconic Sears Catalog. They were done selling through the mail and would stick with stores.
I was crushed. Not that my ideas were wrong. Oh no, I was much too arrogant for that. I was crushed that I could not get them to see the obvious truth, right in front of their eyes.
One member of the audience that day was a consultant with IBM who asked me to participate in a three day brain storming session with Lowes so I could explain the vision. That meeting only had about twenty people in it so there was an actual conversation.
The team at Lowes got it. They were excited about online customer service and sales and advertising. But they faced a serious political problem – they were in competition with their own stores. I explained that the Internet was the same as the telephone. It was just a tool that could be used to aid in marketing and sales. It didn’t have to compete.
Snap-On Tools, State Farm Insurance, Johnson & Johnson – everywhere I went, they wouldn’t tolerate channel conflict and I kept begging them to see that the Internet did not have to conflict but could support and enhance communication with their customers.
Q: How long would it take for these people to figure it out?
A: About twenty years.
A decade ago, large retailers (some sooner than others) granted the company website full status as a store. It would get the same resources as any other store that was selling as much. That was new; a recognition that the Internet was not a fad and that the people working on the World Wide Web were not ‘playing’ with computers. Ecommerce was real commerce.
At the eMetrics Executive Roundtable, I heard a very forward-thinking company make a statement that I had expected to have come true much sooner. But when a participant from another large retailer on the other side of the table said, “Yes, we’re doing that too,” it rang that resonating bell that said this is a major sea change – a shift in thinking that comes from maturity.
“We’re not an independent store. We’re not doing ecommerce. We are part of the infrastructure. We’re a service organization. It’s a little more complicated than just ZIP codes, but when a sale is made online, the local store gets the credit.”
That’s what I had been trying to get others to understand.
Buy it here, buy it now, or let me order it for you online and we’ll ship it to you.
Please see our huge selection on our website.
Buy it now and pick it up in the store this afternoon.
Buy it on your phone, return it in the store.
It all seemed so obvious in 1995.
3. Transparency Wins
Full disclosure: This part was something I learned from a break-out session at the eMetrics Summit and not in the Executive Roundtable. I suspect this is a little too forward thinking at this point and it may only be an instance, rather than the proof in the pudding.
I wrote about some previous resonating moments from Emetrics Summits past in my recent article Profitable Newspaper: Hold the Presses! I mentioned them in a run-up to the current rendition: Joseph Gordon of the San Diego Union-Tribune told how they had become profitable by using analytics to become transparent in their product development and in their selling of ad space.
The condensed version of that article goes like this: Authors (and management) see a daily dashboard that compares the traffic and engagement each one of them ignites. Journalists who have been on staff for decades are competing with those who were hired last week. Their editorial staff remains vigilant that standards are kept high. After all, if you run multivariate testing to its logical extreme, you get porn and the Sand Diego UT is on the opposite end of that spectrum.
Further, the newspaper owns a local cable channel and their advertisers see all the stats on impressions, time on site, clickthroughs, etc., which all get correlated to the Nielsen overnights and to the ads seen on the web by the same individuals.
There is a sense of “information wants to be free” at work here and “transparency breeds confidence” at work here. People do not want to know how the sausage is made, but they do want to know whether or not and how much horse meat is in it.
Trust your employees to make better products when they can see your customers’ response. Trust your advertisers to be happier to work with you and spend more money when they can see the direct results of their expenditures.
Yes, this all sounds very far-fetched but as Wired Magazine celebrates its 20th birthday, it’s good to know that some people, somewhere, are wringing value out of these principles. Or, as William Gibson likes to say, the future is already here – it’s just not evenly distributed.
So those are my three big takeaways the eMetrics Executive Roundtable. Maybe you’ll be at the next one and we can compare notes.
Oh, and here’s a bonus thought, aspirational perhaps, but altogether hoped for: When privacy laws are enacted, one of the potential benefits will be forcing companies to finally create systems around customer data rather than living in transactional, accounting, behavioral or attitudinal silos.
One can hope.