Phase 2: Under way

Back in the end of March, I posted about successful completion of Phase 1. We had released two very successful products: Groups and Calendar, and they were being used by 26,000 times a day 7,500 people. Two months later, our numbers have doubled: every day now, they’re used 49,000 times by 20,000 people. How do we know? We programmed our apps to ping the server every time they’re opened, so we can analyze usage and improve user engagement.

Here is the overview from the last 10 days, if you’re interested:

mysql> SELECT DATE(time_ping) day, COUNT(DISTINCT udid) users, COUNT(1) pings
    -> FROM calendar_ping WHERE time_ping >= "2011-05-01" GROUP BY day;
+------------+-------+-------+
| day        | users | pings |
+------------+-------+-------+
| 2011-05-28 | 12944 | 27429 |
| 2011-05-29 | 14214 | 30504 |
| 2011-05-30 | 16323 | 36986 |
| 2011-05-31 | 17021 | 39326 |
| 2011-06-01 | 17023 | 39488 |
| 2011-06-02 | 16669 | 39085 |
| 2011-06-03 | 15802 | 36066 |
| 2011-06-04 | 13889 | 30173 |
| 2011-06-05 | 15245 | 34086 |
| 2011-06-06 | 17645 | 39671 |
+------------+-------+-------+
37 rows in set (3 min 59.63 sec)

mysql> SELECT DATE(time_created) day, COUNT(DISTINCT udid) users, COUNT(1) pings
    -> FROM groups_ping WHERE time_created >= "2011-05-01" GROUP BY day;
+------------+-------+-------+
| day        | users | pings |
+------------+-------+-------+
| 2011-05-28 |  4047 |  9418 |
| 2011-05-29 |  3901 |  9379 |
| 2011-05-30 |  4213 |  9653 |
| 2011-05-31 |  4714 | 11118 |
| 2011-06-01 |  4754 | 11298 |
| 2011-06-02 |  4628 | 11071 |
| 2011-06-03 |  4644 | 11213 |
| 2011-06-04 |  4102 |  9566 |
| 2011-06-05 |  4189 |  9911 |
| 2011-06-06 |  4119 |  9514 |
+------------+-------+-------+

 

Improvements

Since then, we’ve been working hard on our next phase — increasing user engagement, improving the user experience, and launching revenue streams.

One of the problems that you can see above is that there is a high churn rate. As of this writing, we have had roughly 100k downloads of Groups and 100k downloads of Calendar. While it’s true that every day, our apps are being used by 1 in 10 users who have ever downloaded it, we should do better. What about the other 9/10 of users? Well, we got a partial answer when we checked out our updates. The Groups app was updated on May 8th, and this led to the following chart:

Updates

Updates

Counting the number of people who upgrade gives us a pretty good estimate of how many people still have the app after they’ve installed it. By May 8th, we had around 80,000 users, and 20,000 users kept the app.

So our user retention rate for Groups was pretty okay. 1 out of 4 people kept our app and got the new upgrade when it came out. Our user engagement was less impressive — they used our app less than once every two days.

Calendar was doing much better. It was also downloaded by around 100,000 people, but it was being used three times as much. It actually became sort of a running game on the team to see which app would outdo the other. When Groups became free, it quickly started catching up to Calendar in terms of downloads, but its user engagement was still lagging behind.

Analytics

Thanks to our pings, we had a whole lot of data to look at. Our first idea was to check out how many users who fired up the app wound up using it on day 2, day 3, and days 7, 14 and 21. We found that most people who kept the app until day 2 liked it and continued using it. The problem, then, was with the new users. The app was not easy enough to start using or not compelling enough.

We knew that by adding social features later, we could increase user engagement, but this wasn’t a day-to-day user engagement problem. This was an issue that called for improving the first impression that our app made with users.

What do users want?

So we spoke with our users. Every day, we’d receive several support emails from Groups users. We decided to add a “PS:” to each email, asking them how they easy they found Groups to engage with when they first launched. Turned out, a lot of people had trouble navigating the interface — especially those who weren’t very tech savvy or just got their iPhone recently. Additionally, we felt that if the user didn’t make any groups, they wouldn’t find the app very useful as a contact manager.

The answer was clear — make a tutorial. So one of the things we added to Groups 1.6, our latest version, is a tutorial that you can launch if you haven’t made any groups yet. It walked you through the major 3 or 4 screens of the app, and showed you the main things you can do with it. We just submitted it to Apple, and once we have some data, I will write about whether that made a difference.

Poking Around

The current phase we’re in is all about metrics. If we are going to improve something, we have to be able to measure it. So we’re building tools to better analyze our user engagement, user retention, and user ramp-up metrics from the data we’ve been getting. As far as revenue streams, that’s pretty straightforward. Each in-app purchase is roughly independent of another, so the metric is how much revenue each one is bringing in every day.

Profitability

One of the other things we started working on is becoming profitable. If we will bring in more money than our burn rate, we are able to sustain our business indefinitely. This is a good position to be in when negotiating a valuation for the next round of funding in your company. You can sit back and focus on growing the value of your company, and take investment only if it makes sense, rather than out of fear that you’ll run out of money.

Later this month we’ll be releasing a major product: Calendar Pro. Unlike Calendar, it will be a paid app, probably around $1.99 , and in many ways it will be far beyond the various calendar apps that are selling right now for $15 or more. It will have things like:

  • Google and Facebook syncing
  • Weather forecasts
  • Beautiful calendar views and backgrounds
  • Community managed collections of backgrounds
  • Ability to add events just by typing the information

We are looking to see if it makes sense to add any in-app purchases, as they were just announced at the WWDC earlier today. Anyway, we have some plans with this calendar app, but we aren’t going to reveal them just yet.

Next Steps

While our team is working on Groups and Calendar, we’ve had a couple other projects in the works. The team has grown to about 7 people, some working part-time, others full time. I’ve been busy coding, working on a scalable architecture for what is coming. I had said to myself I won’t code much anymore, but this particular thing I’m working on is very important.

Stage 3 will be all about viral applications. Only after we tweak the user engagement and user satisfaction, and we are satisfied with our metrics, will we add viral features. We are lucky that our apps have great discoverability, and that many our users stick around and love them and keep using them. So we can take our time and get things right before continuing.

But before we dare make something viral, we need to scale. I will talk about our scaling architecture in a future post. It is quite different from the typical stuff you find on highscalability.com . What I am working on won’t just be scalable, but distributed.

In short, there are fun things ahead. Right now, we are just working hard to prepare.

2 Comments

What Makes Companies Great

Investors have all sorts of criteria for trying to predict whether a company will become a runaway hit. Here’s ours:

It’s about the experience of the user (individually) and the user base (collectively).

Each person goes through their own personal story of interacting with your product. At each step, they have a certain impression and attitude. Each person is unique, but collectively, their actions create awesome effects. I have a great respect for the power of the crowd. Edward Burnays, who all but started the field of PR, understood that you can change the world by motivating the crowd, and to do that, you need to appeal to individual people.

Social media is the next generation of this. We get to craft products today that can have an incredibly rich and individual interaction with people, and give them a way to connect with the other users of the product. There are a million possibilities, and the products that succeed are the ones that can become viral and engaging. The products that become great are the ones that give users the tools to create even larger social effects, and have an effect on our culture.

How to be confident that you will succeed

Every person that uses your product will go through certain steps in a user flow, including:

A) They somehow find out about your product
B) They get their first initial experience with it.

E) They get engaged and start investing their time into it.

M) They discover other people using the product.

R) They start generating revenue for you

V) You achieve viral growth

X) Your company is successful.

Z) Your company is great.

Given what I said above, the main question I ask when I want to know a company’s chances of succeeding, is:

How will you get a person to go from step A to step R?

Once you are step R, you are fine. Turning revenues into profits can be done by streamlining your business, charging enough to cover your costs. You can reduce your customer acquisition costs to zero by increasing your viral coefficient. And your revenues should be able to help get you to step V.

Once you hit step V, you are in really good shape. This is mostly because once you start growing virally, a lot of people with money and connections will be interested in helping you grow. With so many people and resources on your side, it’s not about you failing anymore. You will expand in new directions. The investors win, you win.

But for me, it’s all about the experience of the user at every step. What will make them want to progress to the next step?

  • Social proof. When a user first encounters your application, will it come recommended by a friend? Even better, will it have an activity that your friend needs you to join, that you would enjoy more together?
  • Self-expression. Make it easy to publish something to your friends. Then you will want to invite them to look at it. YouTube won because of this.
  • Potential reward. The prospect of earning points towards a reward, such as a badge or extra privileges on the service, motivates a lot of people. Just make sure the game mechanics are not cheesy, and show people a taste of what they will get before they put in their time and effort.

There are lots more things you can do. Our products are all distinguished by the fact that we give groups of people the tools to collaborate on a common activity. So you don’t just invite your friends to view what you’re doing, but to participate in it. It combines social proof, self-expression, and — we believe — increases productivity for everyone.

How can we be confident that we will succeed? Because we have investigated the way people will interact with our products, before we even built them. You can do a simple market study with people from all walks of life, and it will tell you a lot about your chances. Here is what you do:

  • For each screen and stage of your product, mock up a realistic image of what the user will see.
  • Have the person imagine interacting with the product.
  • Tell them the story of how they arrived at this screen. Perhaps a friend just invited them and they are at step A. Perhaps they have the opportunity to invite friends to see their work, or abandon it and leave the site.
  • Ask them to literally “click” or “tap” what they would do next, and describe what they are thinking, and what they want.
  • Try different variations of the screens. We found that the word “Save” is better than “Share”, because people are worried about their work being discarded after putting in the effort to create it.
  • Realize that the probabilities multiply. If a person has a high probability of going from A to B, B, to C, but most people don’t want to go from C to D, then you have to fix that before you can claim your business will be successful.

In my opinion, this sort of research is crucial for predicting the success of an internet startup. Of course, your business model should make sense, given the projected viral coefficient from this research. But if you can show that the business model remains profitable for a broad range of inputs, and if you can show through empirical research that a significant proportion of people have gladly taken the steps A – R on paper, then your project is probably worth funding.

At that point, it’s a question of execution and metrics. If you build your product and your metrics are falling below what your research predicted, then you should work on getting them within your business model parameters. You would do this through repeated A/B testing. Finally, when everything checks out, you can publicly launch.

Good luck and may we all be involved with great companies.

Leave a comment

What’s Your Business Model?

Before the internet, most businesses would generate revenue directly from their customers, and invest in marketing to bring in more customers. Only mass media –including newspapers, magazines and TV — was ad-supported. In fact, even AOL made its money using this traditional business model.

The web became a big game changer. Suddenly, you could publish an interface that everyone with a web browser could interact with, anywhere in the world. Inviting other people, and accepting the invitation, became as simple as sending an email, and clicking a link, respectively. Suddenly, any website could become a “mass media” outlet, and sponsor itself with advertising revenue. New business models became possible. Ad networks for the internet appeared.

At its core, every successful company provides an experience where enough visitors want to progress to the next step, and become clients. In another article, we will talk about the psychological and social aspects of this, but here, we are just concerned with the numbers.

The Formula

Let’s derive a general formula that describes how almost all internet businesses make money.

For a business to be profitable, its revenues must exceed its expenses. The fixed costs include software development, user experience design, and the usual overhead of bureaucracy and running a business. The variable expenses include the cost of acquiring a new client, the resources spent in serving this client (bandwidth, hard disk, support), and the resources spent in serving visitors who do not become clients. This can be expressed as a function of n, the number of clients, and time, t:

Profit(n, t) = Revenue(n, t) – Acquisition(n, t) – Serving(n, t) – Visitors(n, t) – Fixed(t)

  • Acquisition(n, t) – the cost of acquiring n clients
  • Serving(n, t) – the cost of serving n clients
  • Visitors(n, t) – the cost of serving the visitors that led to the n clients
  • Fixed(t) – the fixed cost, it amortizes over time.

When you want to start a business, figure out what your costs would be for this formula. I would recommend making an Excel spreadsheet and plugging in some realistic estimates of the revenues and costs. Then — and this is important — do a sensitivity analysis. Perform a Monte Carlo simulation where you perturb the various revenues and costs, and see what the constraints must be in order for you to make a profit. Often you will find that your business model depends on incredible luck or a huge number of clients in order to start generating a profit. It is not viable — yet.

Viral Can Work

One of the greatest game changers on the internet is that the viral coefficient can dramatically lower the your client acquisition costs, and make an otherwise un-viable business model viable.

Without users inviting other users, your acquisition cost would be a fixed cost per client: spend $5 on ads, get 20 visitors, and convert one to a paying client. Thus, if on average only 1 out of r visitors becomes a client, the cost of acquiring n clients would be:

Acquisition(n) = Ads(n • r) + Visitors(n • r)

Typically, the biggest expense here is the advertising — websites try not to spend too much on visitors who haven’t made an account yet. So when your users start to bring more visitors, really good things start to happen. Your advertising costs go down significantly.

But user invitations is not something you have to leave up to chance. By defining your concepts in useful ways, you can analyze your business model scientifically.

The Viral Coefficient

There are many ways one can measure the rate at which users invite other users. Let’s define the viral coefficient for a given week to be:

V = the number of new visitors that was brought to the site during that month by following an invitation link, divided by the total number of visitors at the start of that week.

By defining the viral coefficient this way, we are able to actually measure it at the end of each week. From this definition, one can see that V has compounding effects on your advertising costs, eventually eliminating the need for advertising.

  • Typically, V is small for most websites — perhaps something like 0.1 . If it stays constant, you will get a geometric series: for every 100 people you bring directly using pay-per-click advertising, you will get 10 for free the next month, and 1 the month after that. The waves die out eventually, but the cumulative sum is somewhere near 1 / (1 – V).
  • If V has been >= 1 for several weeks, you don’t need to advertise at all. Your user base is growing exponentially!

Notice that calculating the V required some very simple operations at the end of each week. However, this definition has a direct relationship to your business model, allowing you to see exactly how far your pay-per-click dollars are going.

Making Your Site More Viral

It’s pretty clear by now that the viral coefficient plays a crucial role in the profitability of your business. The bigger it is, the more profit your business can make, and the faster it attract users.

Be careful what you wish for, because all the other variable costs increase with the number of users. This is where the architecture of your back end servers really becomes important. If it’s not designed correctly, your costs will skyrocket quickly and you won’t be able to scale. This has killed quite a few promising sites. Sometimes, steady growth can be a good thing.

Increasing the virality of your site requires measuring it in the first place. By tracking links inside invitations, you can get a more in-depth analysis of how often the invites get accepted. You can take a google-like approach, and do A/B testing among dozens of different types of invitations. Since invitations are shared away from your site, you wouldn’t use the normal Google analytics tools, though. You’ll need special analytics to track invitations.

The platform we are building will allow you to get these types of analytics out of the box, see how your business model is performing, and figure out which steps you should take to improve your profits. Stay tuned.

5 Comments

Phase 1: Successful.

Since January, we’ve accomplished quite a bit. We’ve successfully launched two productivity tools that people seem to love. The vast majority of ratings are 5/5. Our userbase is growing steadily by about 2,000 people a day. And we haven’t done any marketing yet. I wanted to write about what we did right so far, and our decisions along the way.

TL;DR : Groups! has over 30,000 users, is growing by 500 a day without any advertising, is being used 6,000x a day by 2,500 users. Calendar was launched 2 weeks ago, already has over 30,000 users, is in the top 25 apps on the Mac, is being used 20,000x a day by 5,000 users. And these numbers seem to be steadily increasing. Calendar is growing by 1,500 users a day. There were several factors responsible for our success. Now that we have achieved a good level of user engagement, we will begin monetizing the apps and only then adding viral features.

Groups! for iOS was our first product. It was born out of a personal need: there was no built-in way to manage contacts on the iPhone. So we made one. We wanted it to be the most intuitive and effective group management app on the iPhone, so we integrated it with the iPhone as much as we could.

The first version was received well, but we had some hiccups along the way. When we gave it to our friends to use, it turned out that the iPhone doesn’t let you modify groups that are being synced with Exchange. This caused us to spend another week adding fallback functionality to handle groups internally if native groups didn’t work. As a result, we became the only app on the iPhone (that we know of) which can manage groups even when it was synced with exchange.

We first released a paid version for $0.99, to see what the response would be. We got around 50 people a day downloading it. When the free version came out, we got nearly 2,700 downloads on the first day, and over 13,000 downloads the following week:

Sudden Jump Popularity

The wave of new users subsequently leveled off, was clear that free apps had the potential to become popular without any sort of advertising or PR. The main factors at work here seemed to be the app name, icon and description. The biggest factor of all, though, was the fact that users needed something to organize their contacts into groups, and when they typed “Groups” into the App Store on their iPhone, our app came up first. And it continues to do so to this day.

Pricing: Until now, we’ve maintained two apps on the store: “Groups!” and “Groups! Free”, with the latter consistently receiving ~20x as many new users every day. We briefly experimented with various pricing tiers for Groups. We found that the smaller the price, the better reviews we got. In our next update, we will combine everything into a completely free version of groups, and doing all the upgrades using in-app purchases. This is to make a better experience for the user: they first get the free app, and then make the decision to pay us money on their own time. From a business point of view, we have added several revenue streams which should increase our revenue by several multiples — when we have some data on that, I will make another post.

Usage: We have not had a chance to do any PR yet (although we did get some articles written about us, which may be contributing to the traffic), and spent $0 on advertising. Groups is currently getting a combined 500-600 new users a day.

When the user opens our app, we send a ping to our server, so we are able to track the user engagement and other trends on each of our apps. Here is a snapshot of what has been going on so far:

You can see that only 1444 people used the app today. That’s because I just took the screenshot at 1 PM today. By the way, the query I ran was:

SELECT COUNT( DISTINCT udid ) users, COUNT( 1 ) pings, DATE( time_created ) DATE
FROM groups_ping
GROUP BY DATE
LIMIT 50 , 30

As I write this, Groups currently hovers around the 35th position in Productivity on the iTunes App Store.

Calendar for Mac was launched two weeks ago and already has amassed 30,000 users. We did not cross-promote it from Groups. Once again, I think the biggest factor was the name: yes, it’s really called “Calendar”.

But it’s not just the name, icon and description that make it succeed. We saw a need on the Mac and we filled it. You can now access your schedule from your menu bar. Windows had a calendar in the system tray, and the Mac didn’t.

We used the same principles with Calendar, but this time, they worked even better — probably because the Mac App Store is still new, and there is less competition. We made sure the description and the screenshots indicated exactly what people should expect. The first version of Calendar was just a quick app that we put together which used the Apple APIs to interact with the built-in calendar. You had to use iCal to actually manage your events, but you could see them at a glance with our app.

Pricing: Given what we knew about Groups, we expected about 20x more users to download a free app than a $0.99 one. As the person responsible for launching the apps, this time I found myself asking two important questions. Would we rather have $100 or 2,000 users? And the more crucial question: would we rather have $10,000 or 200,000 users? The answer was clear in both cases: the users were more important. We could always upsell them later.

Usage: We now have over 30,000 people download Calendar. This is the third week that it’s been out, and it’s still in the top free Mac apps (among all apps). In fact, in the last few days it’s been inching it way upwards in the top 25:



We’ll see if we can overtake Evernote :) This was a few days ago.

Engagement with the app has been a source of astoundment for us. We pushed the first version out quickly and didn’t put in the ping to the server every time someone opens the calendar. When the second version finally made it out to the store, we discovered that 4,000 people were using the app a total of 20,000 times a day, and that number has been growing since.

Once again, this was taken at 1 PM, which explains why the last row’s numbers are a little more than half of the previous day’s numbers.

Feedback – People seem to love our apps, if we go by the ratings and reviews. Two factors were very important here:

  1. A very prominent link to report bugs / get support. On the Groups app, “Get Support” is right above “Do you like this app? Write a review.” The order in which people encounter things is very important. If they have a problem, you would much rather they write you a support email than leave a review. On the Calendar app, it’s right there in the menu.
  2. We have a reminder that pops up on the Groups after 3 days, and every 3 days thereafter, that says “Do you like this app? Write a review / Later / Cancel”. It’s one of those standard reminders but it gets people who like the app to write a review.
  3. Reply to your support emails quickly and understand what the people are saying. Often, they have similar problems, and you will develop similar solutions to helping them, so keep track of those solutions and you can copy/paste the text if it really answers what they are talking about.
  4. If you were able to help a person in support, and they thank you, ask them to thank you by writing a review :)

Although the above are strategies to help get more people to write positive reviews, we believe our apps are actually very useful. The biggest indicator we have of this, is that — for some people with unusual environment settings, our apps didn’t work. They kept contacting us about it, asking when we will come out with a fix, and telling us that they wish they could be using the app. Even though for about 1% of our users the app was completely unusable, they did not delete it. A couple people waited for weeks. We were working hard on a new version (which among other things, fixes everyone’s issues) and we are about to release it.  But the length of time people waited for the app to be fixed encourages us in our plans. It says we’re making things people want.

What are we going to do next? Now that we have all this organic usage of our apps, we have a very nice space on which to expand and build. We spent the last two months working with our users to polish our apps, give them the features they wanted, track user engagement, and are about to start monetizing them. After we see some success from the monetization efforts, we will move on to adding viral features and combining the apps into one platform. After that, we will probably be ready to raise more money.

17 Comments

What is this all about?

Hi there.

This isn’t your average company. We started Qbix to change the world in a positive way. We want to improve people’s lives through building better tools for them to use. And with better tools come new opportunities — and new markets. For the past few years, we have been working on a new type of technology, that represents a leap in social communication, and helps people be more productive. In January, we got seed funding and opened the company. Since then, we’ve already made some ambitious progress. Today, we launched this blog, to share our story with you.

Are you an entrepreneur? Interested in business? Passionate about innovation? If you’re reading this blog, you might want to subscribe. We will be writing about what has been working for us, what we’ve done, and how we think about things. Look forward to posts about social networks, psychology and collective behavior, business models, scaling, development and architecture. There is a science to some of this, as we have discovered.

Are you an investor? Subscribe to this blog to find out about our progress. We’ve got a talented team that’s passionate about making a difference. If you like what you see, know that this is just the beginning.

 

5 Comments