No Podcast This Week – Instead, here’s a Chatbot!

Delmus Singleton - Born Dec. 28, 1951 Died June 30, 2001I couldn't find an actual picture of Robert's dad, Robert Martin Sr. This is close enough.You guys may not realize this, but we usually record the podcast on Sunday mornings.  And this Sunday happens to be Father’s Day 2017! So, Happy Father’s Day! Something else y’all may not realize is that Robert and I both have lost our fathers. I lost mine when I was 19. Robert lost his a little over 6 years ago. In honor of our father’s we opted to take Sunday off and just chill. And instead of a podcast, you’re getting this little blurb about what I’ve been working on all weekend. If you’re looking to listen to a podcast, check out our Blog page or head over to our Libsyn page that houses all of our past podcasts.

Chatbots

That’s right, chatbots. Specifically, the chatbot for the FavGeek Facebook page. It’s taken me most of Saturday to rewrite the entire backend, but it’s done. And it’s working! I’ve already had an unknowing participant in my tests. And it worked like a charm. Here’s what I did, why I had to redo it twice, and why I’ve learned so much.

Design

The main reason we decided to implement a chatbot on the FavGeek FB Page was so that we didn’t need a human to answer the questions we got through the page. It was mostly pricing questions (that’s usually what we’re answering on the phone, too), and I knew that we could setup a chatbot to do the dirty work for us. And that meant that I wouldn’t have to answer the Facebook Page anymore! I’m super excited about that. You guys don’t even realize how many times I’d have to stop what I was doing and answer questions on the Facebook page. Now, granted that wasn’t how it was all the time, but it was often enough that it became a problem that needed to be solved. Enter chatbots.

When I first made Oliverbot – the FavGeek chatbot – I just sat down and started making things. I used Chatfuel, and it was very easy to step in knowing absolutely nothing about chatbots. In less than a day I had a working, functioning chatbot. It wasn’t polished, and it wasn’t very useful. But it was there, and it was answering questions (sort of).

I done goofed.I didn’t do any research or really think the flow through on that first go around. When I say “the flow” I’m talking about how people usually interact with us through FB messenger. They start off with a question and want an answer. I started off presenting them with various options they didn’t’ need or want. I put an option for hours up there, but they’d just come from the FB page. It has our hours listed. They don’t need that, and no one is going to click on it! Derp #1. There are also limitations to the amount of options I can put on a single block, and instead of finding a better way of presenting all the options, I just plowed forward using a BUNCH of small blocks. It was hard to navigate and aggravating to go through.

I hadn’t thought through the design of the chatbot. I was just writing to put something down. I’d done no homework. I hadn’t really thought about how it would work at all. It was a disaster.

Redesign

I done goofed.One day last week Robert and I had a business meeting with an associate. At some point last week this associate had tested the chatbot at Robert’s request. He said that it was okay, but it was hard to find what he was looking for. I asked him to be more specific. He said he couldn’t find the price to fix his phone screen. I said that was ludicrous. There was a price button right there on the first screen. He said, yes, but that wasn’t what he was looking for. He was looking for the “broken screen” option. I immediately understood. Derp #2. The customers aren’t looking for the price, they’re looking for the problem.

So I rewrote the entire chatbot, problem first, answer second. That’s how our customers naturally found out their answer so that’s how I wrote the chatbot. The first go around, I actually made a woman mad because the chatbot was so poorly designed that she became aggravated interacting with it. I had to shut it down and respond manually. Like I said, disaster.

Marketing

How does this help you and your business? I don’t know. What kind of business do you have? Are you selling pizza? If so, your chatbot may me an online ordering system! People can go to your page, talk to your chatbot, and have a pizza paid for and delivered without ever talking to a clumsy human. Are you selling women’s country apparel? Chatbots can take those kinds of orders, too. Take payment and tell you that you need to ship an item. There’s a lot chatbots can’t do (humans will always need to be in the loop without some MAJOR programming), but the stuff that takes up a lot of time, answering the same questions again and again? Yup, chatbots have got you covered.

By | 2017-06-28T15:21:51+00:00 June 20th, 2017|Post|Comments Off on No Podcast This Week – Instead, here’s a Chatbot!

Marketing Research 101: A Step by Step Process

Marketing Research 101: A Step by Step Process

Marketing, to the average person, can look like either a sort of business that involves levels of betrayal and manipulation on par with Game of Thrones or an ancient magic spell that only be conjured by the most powerful order to gain results. But if you peeked behind the curtain of the world of marketing, you will be almost disappointed to discover that the Wizard is merely a little man with the right tools at his disposal.

Believe it or not, all the skills that are required in good marketing takes is pattern recognition, research, a base knowledge of sociological behaviors.  Below, I am going to list the most basic of processes we undertake when we dissect and problem solve a typical marketing case.

 

Step 1: Define your Problem – Perform an Analysis

The standard marketing student would immediately declare that they would do a SWOT analysis, while the business owners shrug and shake their heads in confusion. So, I am going to break down this particular jargon, so you can gain some sort of understanding of what you are working with.  A SWOT analysis is merely breaking down the concept into four digestible parts to mitigate what is happening.

  • S – means strengths – In this case, is defined by your companies better qualities that can be used to its advantage.
  • W – means weakness – Is defined by the things that can hurt your company that you are able to change.
  • Oopportunities – Are defined by the things that weren’t in your control but could be used for the benefit of the company.
  • Tthreats – Threats are defined by the things that aren’t in control that can hurt your company.

Marketers tend to develop a sense of strategy after looking at these four factors because it more or less serves as an outline of your particular case, and it is easy to pinpoint a problem there.

Step 2: Develop a Research Plan

Once you have the basic analysis,  marketers usually at this point create a hypothesis based on the conditions of the SWOT. How you create it, comes from asking the right questions:

 Is your business not getting sales because there are fewer visitors at your store?

Is your industry type succeeding or failing? Why?

How can that be fixed?

Are there plenty of visitors but your products aren’t selling? Why?

Who are the return customers and why do they shop at your store?

Why do you have foot traffic but no sales?

These questions scrape only the surface of what a market analysis looks like.

Each business has different problems, different opportunities, different goals, and different strengths. There is no one size fits all when it comes to branding yourself and that is the ultimate challenge of marketing.  It takes an inquisitive mind and a keen eye for detail to ask the right questions and make the right hypothesis for each case.  It’s less like magic and more like detective work.

 

We have to go deeper…

Step 3: Collect the Data

Once you figure out what you have and what you don’t have and how you can adjust, you want to figure out who can benefit from your product and service the most, as well as their motivations, so you can find more people like your current customers.

But it isn’t as easy as it sounds. For instance, once you find out who your target is, if you don’t know how they communicate, you are not going to get the information you need. Not everyone uses a computer, not everyone answers the door, and not everyone answers their phone.

There is already some statistical data about the response rate of certain target demographics like age, gender, and education level, and it is a good place to start but the data represents only the majority of the demographics. That means that there are exceptions to every rule and not everything is set in stone when it comes to data.

Some marketers tend to use more than one form of data gathering to get more results,  such as one member of the marketing group is hosting a quiz for Facebook, combined with someone else from the group knocking on each door in the neighborhood.

Every avenue has its benefits and its drawbacks so be careful when looking for the right data gathering method.

Step 4: Interpret the Information

Once the data is collected, it takes the clearest of minds to mine through the data that has been collected by the entire team, and that is where accuracy is of the utmost importance.  Did a person not fill out the quiz on Facebook all the way? Did they at least fill out the information relevant to your study? If they didn’t do at the very least the last little bit, then it doesn’t count towards your data pool.

Not only do you have to mine the data and compile it as a marketer, but you also have to draw up a presentation, and compile it in a way that the reader can understand so that they can understand what they need to do to change their current situation.

If your reader is a client who hasn’t read past the 11th-grade level in quite some time, you want to avoid special jargon and make the visual data just easier for your client to understand for the sake of transparency between the marketer and the client.

If your reader is your marketing professor, you use every single word in your marketing vocabulary and wrap it up in a proper paper with the best grammar known to man.

Step 5: Draw a Conclusion

Once you have the data compiled and interpreted to your target audience, you recommend what needs to be done and leave them to fix the problem. If they need additional help from time to time but are unfamiliar with what they need to do, it’s okay to reach out and offer a helping hand. In fact, I do encourage a follow up a month from your presentation, just so you can see the fruition of all your hard work, and make a few adjustments.  But the minute that their successes and failures hinge on you entirely and long after you made your case and presented the steps they should take, you leave immediately and work on your next case, unless you are working in-house for a major company.

But the minute that their successes and failures hinge on you entirely and long after you made your case and presented the steps they should take, you leave immediately and work on your next case, unless you are working in-house for a major company.

 

By | 2017-06-28T15:20:26+00:00 June 6th, 2017|Post|Comments Off on Marketing Research 101: A Step by Step Process