Monday, September 9, 2013

Zero To Product: Narrowing My Situational Segment With Analytics

This is part 5 of a retrospective on vizipres: A product I began and worked on for a few months. 

Testing The Real World

I'm moving along in my process of creating my next product, vizipres. I've got a problem space, I've defined some situational segments, I've talked with customers and done some modeling and I have a UI concept I like. Before creating an MVP,  I wanted to kick off another experiment so I could get more information about the situational segment I'm targeting: People who want to organize and share information in order to entertain, educate and influence others. I chose to start this experiment now because it will take some time to get results and in the meantime, I can build out the MVP.

The few things I wanted to learn with this promo site are:
  • More about who my competition is
  • How much money my competition is spending
  • Can I narrow my situational segment
  • Which situational segment has the most demand
I would do this by putting up promo sites, hooking up AdWords and start collecting email addresses. What I'm looking for are:
  • How many people landed on each promo page and interact with it for more than a few seconds (engagement)
  • How many people would sign up to join the pre-release (retention)
The first thing was to put together a simple promo site template.

A Promo Site Template

There's a reason I call it a promo site template - I plan on publishing a few at once, each with slightly different copy and images. After publishing these sites online, I create different AdWords campaigns with each one taking the user to a different promotional page. Again, each promotional page is targeting a different situational segment.

For example, when someone clicks on an AdWord ad associated with 'powerpoint alternative', it will take them to the site which has copy and images targeting people looking for powerpoint alternatives.

When choosing the design for my promo site, I wanted something simple, fast, worked across devices and I could easily swap out copy and images.  The design, copy and images I created can be seen below:

For the copy, I started with something generic with which I could easily swap out different versions. For the layout, I wanted something that was easy to scroll through. Even though a scroll down page can actually do worse, I felt that with a clean design, direct copy and simple images, I could keep viewers engaged. The images I used would be screenshots from my spike - another great side effect of creating a spike. To collect email addresses I simply created a google form and connected it to a spreadsheet. Whenever someone would sign up for the pre-release, I would simply contact them by hand. If I ended up getting more signups than I could handle, then I would write a program to automate the sign up process for me.

Here are a few screenshots of changing the copy when targeting different situational segments:

Managers within a company who may need to present
team / product updates or lead initiatives.

A generic version.

Entrepreneurs who want to convinces others to invest
or join them in a new venture. 

Educators or industry professionals who want
to spread their influence & knowledge.
I would also make some other changes to each template (such as segment specific images), but not too much. I was concerned that too much difference between the versions might make it hard to determine what, exactly, the visitors are responding to.

With some templates created, it's time to see what I can get with AdWords.

Researching Competition & Demand With AdWords

Seeing web searches and average cost per click

I decided to spend a few hundred dollars on AdWords - money well spent in my experience. In the screenshots above, you can see the results of my searches. The one on the right, status report software, is another segment I considered researching. I got this idea from two interviews I had done. One interview was with a sales manager at a inventory research company and the other was a marketing manager at a time tracking SAAS company.

Customer 1 Customer 2
Sales research SAAS Time tracking SAAS
Sales Manager Marketing Manager
Hierarchical company Self directed teams
Problems reporting up chain of command Problems keeping teams up to date

I'll once again point out why it's better to think in terms of situational segments and NOT in traditional customer segmentation. Those last two customers who I interviewed could not be more different; however, they both shared the problem of sharing information within the company.

Another Consideration - Google Trends

Google trends can also be helpful in learning about your target segments. Whenever I am about to run some keywords, I also check out the trends. After running slide show software and powerpoint alternative,  as can be seen below, I can make a few hypotheses:
  • Demand for slide show apps and services may becoming less relevant
  • Searches for powerpoint alternatives seem to peak in the months of April and October / November - possibly due to the need to give quarterly / semiannual reports....
I haven't had enough experience with Google trends to see how the results correlate with product consumption, but my guess is that there is at least a moderate correlation.

Putting It Together And Adding Some Heat

Here's where I am so far:
  • I want to start getting some real world data when it comes to demand and situational segments.
  • I've created a promo site template and created a few site variations - each one with screenshots and copy that target a segment.
  • When these sites are live, I'll be looking for sign ups for the pre-release program as well as engagement within the page
  • I've researched AdWords and Google trends to see what the current competition landscape is and how much $$ is being spent to reach customers.
Just before launch, I want to add another tool to the bag to help me: a heatmap.

An example of a Scrollmap
I used Crazy Egg to track and create Scrollmaps for me. Unfortunately I canceled my account about 7 months ago when my trial ran out and I didn't keep any of the heatmaps; however, the results weren't compelling. The only thing I learned was that most people don't spend much time past the fold ( no surprise ) and those that did generally spent most of the time around where the Steve Ballmer quote was ( check out the screenshot again to see what I mean ).

With my ingredients in the pan, I started to cook .... er launch. After I put up the sites and tested them ( e.g. scrollmap, URL directs for the AdWord testing and the sign up form ) I fired up the AdWord campaigns and waited for the results.....if any ;)

While my sites were cooking, it was time to start my MVP. To do that, I needed take what I learned from my spike and pick my tech stack. I'll talk about that in part 6.