Startup diary: User engagement can prove vital, but tracking it is far from an easy task

Identifying which metrics to monitor may require careful thought from startups.  If your startup is an online service, you'll very quickly run into the term 'user engagement'.

 

It's one of the key metrics that investors look for, and it's going to be a critical number for us here at Voxgig to track. However, user engagement is not a single number, simple to measure. You can easily end up tracking the wrong thing.  What is user engagement supposed to tell you? It is a predictor of growth and future revenue. That's why it's so interesting. It's no good getting lots of people to sign up to your website if they never do anything afterwards. Getting sign-ups is only the first step. If they don't engage because there's no value, you haven't really achieved anything. People's time and attention is valuable, and they have to decide to spend it on your website. That's before they get out their wallets!

If users do find your service valuable, even if they are only on the free version, then they'll keep using it. And as you provide more value and features, and as they find new ways to use your service, you'll start to provide enough value to justify payment - which is the ultimate goal.  User engagement is often tracked with a metric known as 'daily active users'. This metric tells you how many people use your website each day. For many sites, this is a good proxy for user engagement; if people keep coming back, you're doing great. For a daily newspaper, say, this metric correlates quite directly with how much user attention is being captured.

But consider the case of Airbnb. How often does an individual user book a room or apartment with Airbnb? Once or twice a year? In this case, daily active users doesn't tell you as much. Monthly active users is probably more interesting.  Even then, seasonal fluctuations make it less useful than you may think. What does Airbnb actually measure to understand engagement?  According to its engineering team, it uses a combination of individual metrics (such as searches, new listings, etc), and enhances predictive accuracy with a little machine learning. In other words, even that firm finds it hard!

One way to tackle this challenge, and the way we will do it at Voxgig (metrics to be published here in due course), is to identify key repeated actions.  If users perform these actions, then it means they are getting value. It's no good to Paddy Power if users are just browsing its website. It needs users to place bets -that's the engagement that counts.  It's hard to predict in advance, especially with a new service, which actions are really predictive of your users experiencing value. Although you can reason out some of the more obvious ones, it's better not to make too many assumptions.  

For us, we definitely want people to be using our system for event planning and management. That means activity associated with weekly or monthly events is a key indicator of user engagement.  If we get a user, and they just set up one event, and never do anything else, well, that's barely a user at all. But we don't know if a fully planned-out event calendar for the next six months is more or less predictive of future user behaviour (in particular, having them stick around and end up on the paid plans) than detailed and focused use of our system for just a few upcoming events.  

What about sharing event plans with colleagues, or inviting those same colleagues into the system? We just don't know.  Part of the technical work that we are doing at the moment is preparing to measure all these activities. We want to keep an open mind so that the data can tell us what counts and what doesn't.  Once we've identified metrics that are good predictors of user engagement, these will be very useful to focus our use of our limited startup resources. There's no point deepening features that don't provide value.  

One problem we'll face is the volume of data, not just in terms of data points, but individual metrics. Luckily, these days, there is an entire eco-system of software-as-a-service options. The business model is big enough now, particularly as more and more traditional industries adopt it, that almost any SaaS business need can be solved (for a price). Handling all this data is not something you should attempt to do yourself. Not only is it a waste of your software developer's time, but you won't build anything as accurate and reliable as the specialist services.  Just to make a dangerous prediction, we feel that sharing event attendance and planning calendars with colleagues is going to be a key metric for us.  

Synchronising activities is always a challenge, and we do make it easier. We made a decision to invest developer time in building our features to support this use case. Soon, the gods of user engagement will give their judgement.  If your startup is an online service, you'll very quickly run into the term 'user engagement'. It's one of the key metrics that investors look for, and it's going to be a critical number for us here at Voxgig to track. However, user engagement is not a single number, simple to measure. You can easily end up tracking the wrong thing.  

What is user engagement supposed to tell you? It is a predictor of growth and future revenue. That's why it's so interesting. It's no good getting lots of people to sign up to your website if they never do anything afterwards. Getting sign-ups is only the first step. If they don't engage because there's no value, you haven't really achieved anything. People's time and attention is valuable, and they have to decide to spend it on your website. That's before they get out their wallets!

If users do find your service valuable, even if they are only on the free version, then they'll keep using it. And as you provide more value and features, and as they find new ways to use your service, you'll start to provide enough value to justify payment - which is the ultimate goal.  User engagement is often tracked with a metric known as 'daily active users'. This metric tells you how many people use your website each day. For many sites, this is a good proxy for user engagement; if people keep coming back, you're doing great. For a daily newspaper, say, this metric correlates quite directly with how much user attention is being captured.  

But consider the case of Airbnb. How often does an individual user book a room or apartment with Airbnb? Once or twice a year? In this case, daily active users doesn't tell you as much. Monthly active users is probably more interesting.  Even then, seasonal fluctuations make it less useful than you may think. What does Airbnb actually measure to understand engagement?

According to its engineering team, it uses a combination of individual metrics (such as searches, new listings, etc), and enhances predictive accuracy with a little machine learning. In other words, even that firm finds it hard.  One way to tackle this challenge, and the way we will do it at Voxgig (metrics to be published here in due course), is to identify key repeated actions.  If users perform these actions, then it means they are getting value. It's no good to Paddy Power if users are just browsing its website. It needs users to place bets -that's the engagement that counts.

It's hard to predict in advance, especially with a new service, which actions are really predictive of your users experiencing value. Although you can reason out some of the more obvious ones, it's better not to make too many assumptions.  For us, we definitely want people to be using our system for event planning and management. That means activity associated with weekly or monthly events is a key indicator of user engagement.  

If we get a user, and they just set up one event, and never do anything else, well, that's barely a user at all. But we don't know if a fully planned-out event calendar for the next six months is more or less predictive of future user behaviour (in particular, having them stick around and end up on the paid plans) than detailed and focused use of our system for just a few upcoming events.  What about sharing event plans with colleagues, or inviting those same colleagues into the system? We just don't know.

Part of the technical work that we are doing at the moment is preparing to measure all these activities. We want to keep an open mind so that the data can tell us what counts and what doesn't.  Once we've identified metrics that are good predictors of user engagement, these will be very useful to focus our use of our limited startup resources. There's no point deepening features that don't provide value.

One problem we'll face is the volume of data, not just in terms of data points, but individual metrics. Luckily, these days, there is an entire eco-system of software-as-a-service options. The business model is big enough now, particularly as more and more traditional industries adopt it, that almost any SaaS business need can be solved (for a price). Handling all this data is not something you should attempt to do yourself. Not only is it a waste of your software developer's time, but you won't build anything as accurate and reliable as the specialist services.

Just to make a dangerous prediction, we feel that sharing event attendance and planning calendars with colleagues is going to be a key metric for us.  Synchronising activities is always a challenge, and we do make it easier. We made a decision to invest developer time in building our features to support this use case. Soon, the gods of user engagement will give their judgement.

Source:  Irish Independent

Share

JOIN OUR MAILING LIST

Contact Us

South East BIC CLG.
Unit 1B, Industrial Park
Cork Road, Waterford
X91 DT38

Tel: +353 51 356300

Social Share

Our Partners

 

 

We use cookies to improve our website and your experience when using it. Cookies used for the essential operation of the site have already been set. To find out more about the cookies we use and how to delete them, see our privacy policy.

I accept cookies from this site.

We use cookies to improve our website and your experience when using it. Cookies used for the essential operation of the site have already been set. To find out more about the cookies we use and how to delete them, see our privacy policy.

I accept cookies from this site.