How to be a Twitter Bad Ass #5: Data Mining?

Jim Murray, Mike Spencer, Jeffrey Summers, and Regina Costa inspired this post. They all took exception at my method for Twitter engagement. Jim later said he didn't so much "take exception" as "question" it. Potayto, potahto... Same thing Big Jim.

Regina found it too complex, too time-consuming. She wanted a simpler method.

The guys thought I was just embracing vanity metrics since I couldn't provide an accurate ROI for all this stuff. 

I don't agree with either view, but they are from people whose opinion I value. Whether I agree or not, I must take their points seriously.

That got me to thinking. I revisited the logic that brought me to here.

I had a tough time quantifying Twitter's effect, so I figured I'd just build a big following. The bigger the following, the greater the odds that the right followers would be there.

It's an inelegant solution, but a solution nonetheless.

The method I've used so far has worked. Anyone can implement it if they are willing to put in the time. But, it's a buckshot strategy. I prefer sniper strategies. I even wrote a LinkedIn article about why views suck.

Why I'm so convinced that Twitter works for me

I'm convinced because I see the results. I can't measure where exactly they come from, but they're there. I assume the fault lies in my understanding of the issue, not in the issue itself.

I'm convinced because creating the content is only half the story. Promoting it is the other half. Content, and its cousin, content promotion, are not the stars of the show. They are supporting cast members at best. They are roadies at worst.

Content does not bring in business. It supports those who do. (tweet)

Content does not sell. It informs. (tweet)

Content guides. Content seduces. (tweet)

I won't rehash all that here. I posted enough on that subject in "What if David and Goliath Opened a Jamba Juice Instead?"

I'm no fan of vanity metrics. 

It always bugged me that I couldn't put a finger on what was doing what. It was enough that some thing was doing something. It was worse when clients would ask for details I couldn't provide, it was a lot worse.

It's a little embarrassing to say, "It works because it does, but durned if I know why."

Jim, Jeffrey, and Mike convinced me to take another kick at the can. Turns out I'm glad they did. It was about this time that I started thinking about the complexity issue that Regina brought up.

I looked at ROI from a different angle.

ROI is simple division. How much do we get back for how much we give? It's just a ratio. Increasing the "R," or cutting down the "I" has the same effect. Better yet, we could do both.

I thought I can't pinpoint the Return part of the equation. Why not just reduce the Investment part to the point that it no longer matters?

Simplifying the process

I started digging into the Twitter API in earnest.

API is an acronym for Application Programing Interface. It's a set of commands and methods that one application can use to access another.

Twitter's API is broad, deep and powerful. The more I dug into it the more I dug it. (tweet)

There is a ton of data available for the picking. So much so that I'm sure I may be able to create a true engagement platform for Twitter. 

Don't get too excited, though. It can't be completely automated. That isn't a tech issue. It's a Twitter rules issue.

Twitter frowns on complete automation. In fact, it downright forbids it under threat of account suspension or even cancellation.

That's okay. The Twitter rules allow me to automate nearly everything. I am just not allowed to automate the actual act of following or unfollowing someone. 

Believe it or not, it's okay to provide follow-links to click. It's forbidden to just go ahead and follow.

I guess they have to make it that way to avoid server-based transactions. I can click 183 times in a minute without forcing myself. (Yes, I just timed it.) That's glacial in server terms.

A server can easily fire up thousands of actions per minute. Some servers can do thousands per second. I understand why Twitter does it that way. But, still, it's a shame that they have to limit legitimate functionality to avoid spammers.

There may be a grey area at work here. If anyone from Twitter reads this, maybe they can enlighten me.

I have no hard data to support the following allegation. I bet that the sheer volume of Twitter data causes many users to avoid it. It doesn't need to be so.(tweet)

It's what Trent Selbrede calls, "noise."

That's too bad. It's that crazy volume of data that makes it wonderful. It's just that we humans aren't well suited to absorb it. We need servers.

Based on what I found in the API I can target who to follow with great precision.

• I can focus on those people who share my interests and are active in the community.
• I can talk to people who are listening. I can talk to them when they are listening.
• I can trace activity and engagement back to where it began. That's the beginning of measured ROI.

In short, I can target my audience to those who want to engage and with whom I want to engage. That's the final goal we all share. 

We all want people who will digest what we offer. We want them to react to it. We want them to act on it. 

We want full engagement. We want that full engagement as a high ratio of total attempts at engagement. That means we need to target tightly.

On a platform like Twitter, or Instagram for that matter, the high volume of data precludes any attempt to control it manually. Only automation can save our heinies.

This will likely be the last post in this series. 

The engagement platform is just about finished. Time permitting, I will start testing sometime next week. It's a rather complicated synthesis of four programming languages. It needs web and database servers to work.

It isn't something you can just toss together in a jiffy, but it's been a fun ride.

• Step 1: Assemble data from the massive Twitter stream DONE
• Step 2: Analyse that data. DONE
• Step 3: Find actionable patterns DONE
• Step 4: Act on those patterns DONE
• Step 5: Test, measure and repeat 

With a little luck, it will be an elegant solution to the Twitter ROI dilemma. I just might be able to figure out a meaningful path to Twitter engagement.

If so, I'll post a follow-up.

Wish me luck.

Ghosts that Write Hive: Get Writing Help from Pros

Follow me on Twitter

Follow me on LinkedIn

Follow me on 

beBee