Thursday, 27 July 2017

Getting email up and running for side projects

Photo by Songeunyoung on Unsplash
Rectifying my previous mistake - not building community by using email engagement - has proved to be a learning experience! Since our investment is minimal (and we have no real users yet!) the tools that we use all come with some constraint/trade-off. This next bit explain this is going to be slightly technical ...

With Heroku it has surfaced as putting in CNAME records in the root of the domain. With register.ly this then stops us adding any other root records, for example mail servers. Cloudfare has a service that will flatten CNAME records for you into A records so everything plays fits together. So a switch of nameservers and 48 hours later we had email in place.

The services that we have chosen for email are Zoho and MailerLite. Zoho has a Office365/GSuite vibe going on and includes a mail list server. This setup not only works for newsletters but also the transactional emails. So we get account welcome, platform updates, and follow up emails as well. As someone who has been creating workflow systems for nearly two decades I am impressed by their automation setup. The main problem has been restricting myself to a very simple MVP. Too tempting to build a baroque workflow straight away! So this gives us the tools to get a connection to users and the app, with community building. I hadn't considered email lists, but have now added that to the roadmap.

As this project is taking shape we are using more of the skills from our professional lives. For example, a lot of the code to integrate these services now lives in a GitLab kanban board. We are creating issues for specific bugs instead of general cards in Trello then pushing code changes to git. Not sure why this is. Could be the launch approaching has caused it to kick in? Or it's getting to grips with DNS has reawaken what it's like to work in an early stage company.

As we are getting ready to launch we have also added a system status page. UptimeRobot has a great system for doing this. It evens includes the obligatory Slack integration for alerting. This didn't use much technical skill, add a DNS record and it worked. It was funny how excited setting this up made me!

I did originally write a whole post on the onboarding process. This fell a little flat as it was missing a lesson or stand out experience. The aim is to have more to share on that next time.

Saturday, 8 July 2017

Building awareness of Bashfully

So to validate an idea you need to have people be aware of you. The first step in Introducing bashfully was a Twitter account. This provided someway of people finding out about project, promoting the survey, and getting to the landing page.

Publishing the survey was one of the first pieces that I had available to promote. It was also vital in starting to get some evidence for the problem that we were trying to solve and the audience.

The landing page was a very important piece in getting started. When this was created it gave us a way of sharing progress, to explain the concept in more detail. This then meant that when we directed people to the survey from the landing page we could get better feedback.

Another key feature on the landing page was a sharing button on registration, this gave people who were interested a way of generating some word of mouth interest. Around the same time a friend asked me a question about her website and I noticed some interesting meta tags. This took me on a journey learning about Twitter cards and the Facebook Open Graph template

So, after a bit of experimentation I came up with this for our landing page

<meta name="twitter:card" content="summary" /> 
<meta name="twitter:site" content="@bashfullyapp" /> 
<meta property="og:url" content="http://bashful.ly" /> 
<meta property="og:title" content="Do you want a better online resume?" /> 
<meta property="og:description" content="Tell the world about your achievements through storytelling on Bashful.ly!" /> 
<meta property="og:image" content="https://bashful.ly/images/hamster-background.jpg" />

Which when you copy a bashfully link to Twitter gives you this ...



As the functionality develops we will be exploring this further! I wasn't expecting to learn this kind of thing when we started this side project, but it certainly has given me an extra insight to social media for my day job.

Once the registration process, landing page and survey were all up and running I started sharing. First with HBX networking groups as they are part of the target demographic. Then on twitter, first just from the bashfully account, later from my personal twitter and LinkedIn accounts.

Another important factor was starting to write blog posts and share on personal social media to boost awareness. The blog that really got interest going was one about the technical infrastructure. This lead to a lot of interest from friends and acquaintances. I'm guessing because of their background and the subject matter made it seem more "real". The lesson here is to write a piece that will resonate with an audience, then promote it to them.

As I was doing this and collecting together research I realised a big mistake. I hadn't thought about emails sharing useful information to people who had registered. This would have served two aims, the first the reason people were interested in the product in the first place - information and guidance about creating profiles and selling their skills. The second, more selfish reason, was to create engagement. Make sure that they remembered us when we were ready for the public beta.

It's a bit late to do that now as we are making such good progress! But it will come in soon after we launch. One benefit of doing it this way is that we will have a working site people can use and more concrete news that we can share.

Coming next are the lessons learned in building the onboarding process...


Monday, 3 July 2017

Chromebook on the go after 3 months

Bit of an update on cloud working and the tools that I'd picked out to investigate over the past three months: Draw.io is a great alternative to Visio, although working in a Microsoft dev and productivity stack I'd still stick with Visio. It's missing some extra collaboration features that would give it the edge. Which brings me onto Mockflow. This is a great wire-framing tool. The collaborative element allowed for some "paired designing". Each with our own laptop to look things up while the other added elements that we'd agreed to.

Pixlr editor is a good replacement for what I need in an image editing tool. Haven't missed the desktop aspect. Similar experience with Caret, although I haven't needed a text editor that much. Office online is a bit frustrating as it loses what should be a key feature - consistency. All I want to do is put some bullet points in!! But it suffers from the same issues that ALL WYSIWY HTML tools I have ever used suffer from. It gets confused between where the formatting applies.

As a laptop I have no complaints with the Chromebook, it makes the hardware a commodity item. One that I don't have to worry about. The ability to mount Dropbox and OneDrive as file stores also negates vendor lock in. I can have multiple back ups and work where it is appropriate. iOS even comes with a Google Drive widget so I can email my mock-up files on the go!

Where it gets interesting is with Codeanywhere and Cloud9. I prefer Cloud9 as being a bit more user friendly and a tad more RAM in the free pricing tier. It took no time at all to get a virtual machine running Linux and install R and create a PDF containing a graph.

This does open up all kinds of possibilities. For hobbyists, you can try out a new stack without much knowledge about the setup or ops. Then if you don't like it get rid of and you don't have clutter on your system. Then have a clash of libraries the next time that you want to try something. It also provides the ability to get some experience while using similar technology to enterprise systems .

For startups you get access to enterprise style infrastructure, if not scale. One of the most surprising things for me as I've looked at a couple of side projects is just how easy it is. I'm sure that for $9/month you could host an email marketing system for an SME. It'd (optimistically ;) only take a few weeks of full time work to get it up and running.

For enterprises that last sentence may seem worrying. How can competitor analysis keep track of specialist companies? Especially if they can pop up within a month and steal your business? Well, by doing the same thing. Get your pioneers to experiment. No need to wait for ops to be available. No need to worry about too much in fact. Do the experiment. Learn the lessons. Then either bring the experience in house or do some knowledge transfer. Or migrate to your usual infrastructure and tech stack.

The other aspect is that with scale in customers, you also have an advantage in problem knowledge. One way to keep customers is to double down on understanding their problems. Then solve them. Also keep revisiting this, to keep actual problems at the forefront.


* my optimistic and purely hypothetical SME marketing tool:
  1. Get the sending side sorted with basic metrics recorded (this would be heavily manual driven process by the startup)
  2. Get some reporting based on metrics for performance of first send to users (first time you give access to system)
  3. Get some basic content management for next send
  4. Get billing in place ready for first invoice
  5. Then refine each aspect of above each time you sell it

Of course this is dependent on experience with the technology, free time, and understanding customer you could get each stage above down to about a week. It's also not going to be competing with Pure360 after a month, but it will do a job for someone.

Sunday, 25 June 2017

Building Bashfully - a brief background

The first guest post here on "Part of the Process" from Martyn Osborne as he explains the Infrastructure side of bashfully, a parallel concern to testing the vision.

Since we've recently publicly unveiled Bashfully, it seems prudent to run through the technology stack and why we chose it.  I’m hoping it may also at least partially explain why starting a side project was appealing to me!

Backend stack

Naturally, there were some elements I wanted to focus on when selecting the language:
  • Fun.  As it's a side project (in addition to a full-time development job), I wanted something interesting to play with in my free time.
  • Different.  During the day I spend most of my time in C#, F# and JavaScript in Windows-land.
  • Functional.  I love F# and the paradigm, so that was the direction I wanted to go.
  • Web sockets.  This project is a toy - and I wanted to play with them at some point (of course, when appropriate!)
  • Practical.  Yes, I wanted a toy.  I also wanted a project that will eventually materialise!

I had been keeping a close eye on Elixir for quite a while, and Bashfully seemed an excellent excuse to use it in anger!  Elixir also has an excellent web framework with Phoenix, and an exceptionally helpful community.

So far, Elixir genuinely makes me happy.  I personally think that’s high praise, and I may dive into more detail there in the future.

PostgreSQL suffices for persistence.  Elixir has good support with Ecto, it’s easy to run on my machine, and Heroku allows remote PSQL sessions.

Frontend stack

To be honest, I battled with the idea of making Bashfully an SPA for a while (and I have some expertise in that area).  However, I now believe that would have been overkill; especially at what is essentially a prototype phase.

Instead, we’re primarily using Phoenix to render pages on the server side with Vue sprinkled throughout where advantageous.  I’m not particularly attached to Vue yet, but it’s sufficient and I am eager to explore further.

Hosting
                                                        
On the hosting side of things, Heroku has been fantastic.  We’re using GitLab CI to build, test and immediately deploy Bashfully to our Heroku staging environment;  we can then promote that build to production with Heroku’s pipeline feature.  This was working well within an hour!

Even though this setup was primarily to make my own life easier, it also had the interesting effect of allowing Neil to easily make design, wording and content changes to the application (via GitLab’s excellent editing UI) and have them deploy to staging without any technical barriers.

Security

This bit is pretty important.  We have worked on a similar web application for a while and have learnt several lessons along the way; these lessons have been incorporated into Bashfully.

We decided from the very beginning to go passwordless and use third-party authentication providers (Google, Facebook, etc).  This was trivial to set up (we’re using Ueberauth) and means we don’t have to process or store credentials - very helpful in the age of data breaches!

Site-wide HTTPS has also been implemented courtesy of LetsEncrypt via Heroku.  

My advice

For anyone else attempting a side project (and with the intention of going live in a reasonable timeframe!), I would suggest:
  • Don’t pick any technologies too esoteric.  Make sure they’re fun, but also practical!
  • Do pick some technologies that you are familiar with.  Writing good quality idiomatic code when learning one technology is hard enough - don’t try it with too many.
  • Keep the infrastructure simple.  I toyed with the idea of building the infrastructure myself on AWS;  but not doing so let us concentrate on the mockups, journeys, and the application itself!
  • Define your vision early on.  Staying focused on building the right things is critical, and a vision helps (especially as a developer!).
  • Break the project up into smaller chunks.  Something like Trello is a great help.  I would argue that this is a critical skill for any developer and one that a side project can help with.

Thanks


If you made it this far - kudos!  I will probably be expanding on these topics in the future, so keep an eye on this blog!

To add to this, not only was I using the GitLab UI but I have created, merged and squashed feature branches. I have even wired up some page handlers unaided without knowing any Elixir or Phoneix!

Thursday, 22 June 2017

MEETUP: "A look into what every Product Manager forgets " at ProductTank Brighton

Three varied viewpoints and interesting talks at tonight's ProductTank, around what Product Managers usually forget. First up was Mark Rodgers sharing how they overlooked things in the first iteration of the new image search functionality at Brandwatch. I can completely relate the situation in my own work. That thing where when you see the product with real data and a real usage situation you suddenly notice something and think "how did I miss that?!", since it now seems so obvious. This made me feel a bit better that if someone, as experienced with Mark and with his team at Brandwatch, can make that mistake it's not surprising I do.

Next was Ben Sauer from ClearLeft. His talk was a more abstract look at companies culture. For example, how it is all around us but that we don't notice it. And consequently (or maybe because of?) we don't discuss it enough. He recommended a couple of books to read Creativity, inc  by Ed Catmull and 
Nonviolent Communication: A Language of Life by Marshall Rosenberg that I am looking forward to reading. Especially as he said the lessons of applying NVC to his life had improved his relationships. One particular piece of advice that stuck with me was to use comedians for change. That is if you are using videoed user feedback to convince "higher ups" that something is a bad idea, get someone funny. He put forward the suggestion that it helps if people can laugh at themselves when admitting a coursed of action was wrong.
The final talk was Tim Stamp from Rakuten talking about security and how this can damage business reputation when done badly. I know that I have thought about user permissions a step too late in the design process before. One piece of advice he gave was to get someone in with a security focus when discussing the user journeys so that they can suggest how attackers could abuse the system. A related point that Ben made in the panel Q&A after, was that we don't have discussions about security/usability tradeoff enough. So get security in at the user journey mapping should help that happen as well.

It was also very satisfying listen to Tim talk about passwords, and if possible avoiding them by using single sign-on. Exactly the approach, and for exactly the same reasons, why we have avoided them on my side project!

Sunday, 18 June 2017

Introducing bashfully

So in my last post on the "untitled side project" I said that I would elaborate more on the idea. When we started the project we had a strong idea of the kind of project that we wanted to do. And after settling in the area I started doing some reading around and thinking about who we could help. What change were we looking to make in the world? How were we looking to alter behaviour? (a litmus test for innovation) 

The project is called Bashfully. Partially a play on the concept of being bashful and shy about talking about your achievements. Partially because the name could be part of the domain in the .ly TLD ;-) This project is to provide an online resume to help people shine in ways that the current sites do not. To help them be proud of what they have achieved. Finally, to allow them to see their career development as a journey and set their own narrative.

The Trello card entitled "what are we hoping to learn?" says:


MVP is to test the hypothesis that there is a market for a resume site and network for people in a creative role or within 3-4 years out of university. These people are not currently served by either LinkedIn because the format is quite restrictive/unispiring or Facebook because it is too informal and basic. Although both sites offer good networking and the target demographic are already somewhat established in at least one of these.
As the first step to sense check this I created a survey and shared on social media. Around the same time, we created a landing page. This had two purposes, 

  1. to share awareness of the project, including the survey 
  2. to build up the functionality iteratively
The first part of this functionality was the registration of interest. This is the foundation of the user authentication that the full site will use. It allowed us to explore the different APIs that we wanted to use and actually integrate with them.

The initial results from the survey were great. One idea suggested was almost word for word how I would describe it - with the caveat we need to be careful of confirmation bias! Another respondent gave us an idea that was brilliant and we hadn't thought of. 

The output of looking at these survey results also helped firm up the vision from the initial hypothesis.This was a vital thing to get in place. Any project and especially product needs to have a clear sense of:
  1. Who you are helping?
  2. Why you are helping them?
  3. How will this help them? What is the positive outcome for them?
Without this, any sense of prioritisation or ROI calculations that you will hope to do are likely to be a mirage at worst. At best you'll have a useful proxy, but that can change without you noticing. This is as true for a 2 person side project as it is for any commercial organisation. Probably more since we are doing this for fun and not profit, and launching is part of the fun!

Thanks for reading. The next instalment is likely to be on crafting the user journeys and mockups to test with real people (aka "users").

Further reading


Friday, 16 June 2017

Is "addictive" app design ethical?

Many years ago I used to live with someone who commuted an hour longer than I did. This was way before smartphones, social media, or even cheap laptops. So I used to do the washing up as part of my evening ritual of winding down from work - luckily she was also as relaxed about leaving the washing up as I was.

Thinking back to this time the main method of communicating online was via forums around interests. So Eurogamer, Get Your Boot On, and North Stand Chat for example. Here the conversations are around topics. The nature of content on the sites is discoverable and predictable. I could easily see new information that I was interested in. I could also easily show someone else something interesting later.

Fast forward a decade with smartphones and pretty widespread internet connections I can get a notification ping up on my phone instantly alerting me that someone I know has posted something for the first time in a while. It will probably be a cup of coffee. Do I really need to be that connected?

Don't get me wrong, in my line of work at 15below creating alerts for people on the move, who need really timely information about their journey, can be the difference between missing or catching a flight. The ROI of the joy of making a flight you would otherwise have missed is hard to capture, but I'd certainly call it progress.

What I wonder about is the switch from the topic based forums I used to frequent, to people based "social networks" that are a virtual stream of consciousness. Sure they can be easily discoverable. They can put people in different countries who have never met in real friendships. However, they can also be unpredictable. Feeding that part of our lizard brain that feels rewards for novelty. 

The worst for this is LinkedIn. I almost feel like the second I see something I have to like it, read it or share it. Otherwise, I will never see that post again. This actually makes the experience stressful. If I see multiple items of interest when I open the app it's a race against time to complete them before I move on.

It's a similar story with Twitter or Facebook. Constantly refreshing or switching between searches to see what's "new". Sometimes I even fall into a zombie-like state, swiping and refreshing. When I stop I'm exhausted falling into a daze. This undoubtedly has an impact on real life relationships. With mental health. Is the disruptive nature of the curated timeline worth the ads it allows the platforms to push in our faces? Because social networks are becoming so interwoven into our lives that not being present there can be just as isolating.

My Sunday morning last weekend really brought it home to me. When I got up, instead of the usual coffee and social media routine, I did some housework. I unloaded the dishwasher. Washed up some fragile glasses. Put some washing on the line. When I had finished and sat down with my coffee, I felt much more awake. I was also more relaxed.

As product people do we have an ethical duty to consider the impact of our creations on people? I use the term "people" here deliberately rather than "users". This is probably an enabling term that allows us to distance our actions from the impact.

I will leave you with this thought from the Pope - and whether online or in real life keep your interactions meaningful.
Sorry for the rant. I know it's not overly original. But I feel it is important. We need to be more aware of the issues and avoid them.

Further reading


Getting email up and running for side projects

Photo by Songeunyoung on Unsplash Rectifying my previous mistake - not building community by using email engagement - has proved to...

Highlights