Install insights on the Shopify app store 2021
We use data to learn more about installs on the Shopify app store with general stats and 4 pieces of targeted insights.
Welcome back to the Marketplace Apps newsletter👋 We’ve been busy enriching our datasets and we are now collecting even more data points on the Shopify app store to generate insights from, including:
Install growth (from mid-July 2021)
Revenue (displayed as MRR brackets)
Rank change over time
There’s been a fair bit of work that’s gone into this so big thanks to Lukasz Wiktor and Jakob Greenfeld for their help in making this happen 👏
All that data in addition to the data used to create this report is available on Marketplace Apps. So if you want access to the data to generate your own insights, check us out! 🙌
In this post we are going to focus on Installs and have a look at what we can learn from that set of data points. If you are not familiar with the Shopify App store, I suggest reading the “Shopify 2021 Insights Report” we produced previously. It’ll help you get up to speed with the Shopify app store that will make this post more insightful.
I’ve structured the post in a few parts:
General stats - High-level stats if you want some insights quickly.
4 key insights - More detailed analysis that use those high level stats & more to drive insights including:
Freemium is the superior pricing model for installs
Sourcing & Selling Products is the best category for installs
The Drop-shipping subcategory drives Sourcing & Selling Products install outperformance
Number of reviews is positively correlated with installs but average rating is negatively correlated.
Feedback request: I’m considering starting a weekly/bi-weekly newsletter that highlights 1 - 3 acquisition opportunities using data. If you would be interested please leave a comment to let me know.
Note: We currently have install data for the top 80% of apps by installs. So we are missing install data for the bottom 20% at this time. We are working on this and should have the complete set shortly.
General stats 📈
At the time of writing there are 9.55M installs (from the top 80% of apps by number of installs)
Average installs per app across all apps = 1,972.4
Median installs per app across all apps= 160
Freemium pricing models have the highest average installs and median installs:
Average = 2,944.7
Median = 349
Paid only apps are the worst pricing models in terms of installs with:
Average = 1,271.1
Median = 131
The “Store Design” category has the highest share of installs with 22.3%
The category with the highest average installs per app is “Sourcing and Selling Products” with 4,199.3
The category with the highest median installs is “Conversion” with 192
Installs and number of reviews are positively correlated with a correlation of 0.44
Installs and average rating are very slightly negatively correlated with a correlation of -0.02
1. Freemium models are superior in terms of installs 💡
There are three pricing models that I examined in reference to installs
Free - These apps are completely free
Freemium - Apps that have a free plan and customers scale up
Paid - Apps that only offer paid plans
When we examine installs, Freemium models take up the lion’s share of the installs with 40% of the installs (3.8M), Paid has 34% (3.23M) and Free with 26% (2.5M)
Diving a little bit deeper there is also variation in the median and average number of installs across pricing models.
Most notable is the far higher performance of the Freemium plan relative to the app store overall and the other plans. With a median number of installs of 349 that’s 118% higher than the app store median.
This is surprising as you would expect completely free apps to have the highest average and median install figures. This indicates to me that there are likely a lot of Free apps with very few installs and just a few free apps with a very large share. However, I’d need to dig in a little deeper to confirm.
Nevertheless, based on this data, if you are developing a new app or looking to improve an app you’ve acquired, building in a freemium model is the simplest way to drive growth.
2. Categories - “Sourcing & Selling products” category outperforms 🚀
We can take a look at installs on a category level. It’s better to look at this on a percentage basis as opposed so absolute numbers. This is because currently we have install data for the top 80% of apps by number of installs. We are working on obtaining install data for the remaining 20% as we speak. That said, whilst the absolute numbers will change, I don’t expect the percentage splits to change in a meaningful way.
I have also checked which categories the apps that we don’t yet have install data are located in and they are spread evenly across all categories. So we can be pretty confident in the results.
You can see the percentage split of installs by category on the table below:
The percentage splits likely won’t change significantly for two more reasons:
We’ve captured 80% of the data and the data that contributes the most towards installs. Therefore, it’s unlikely the bottom 20% of apps with the least installs will meaningfully move the data.
The percentage of installs, pretty closely matches the category’s percentage of apps on the app store (see the table below). That indicates that the average installs per app is pretty consistent across categories. We’ll discuss that in more detail below.
There are two categories that are anomalies to this trend. “Marketing” and “Sourcing and selling products”.
As shown on the table above, both categories have significantly higher percentage of installs compared to percentage of apps on the store (look at the variance column). For me, this pushes me to focus on those two categories to targets. As developers and/or acquirers in those categories will be more likely to gain an outsized percentage of install with their apps.
Average installs per category
The chart above shows the average installs per app by category. The average for the app store overall is 1,972.4 and as suspected both “Sourcing & Selling Products” and “Marketing” outperform the other categories in terms of average installs. Sourcing & Selling Products has an average installs of more than double the Shopify store average.
However, as I’ve discussed in previous posts, averages on the Shopify app store can be misleading. A concentrated number of apps make up for the majority of activity on the store and therefore looking at medians is usually more instructive.
Median installs per category
The median installs for an app on Shopify is 160. The “Conversion” category immediately stands out with 192 installs per app with “Store Design” close behind with 179. However, I’ve highlighted “Sourcing & Selling Products” with 173 with the third highest median installs. For me, this is the final data point that completes the story that “Sourcing & Selling Products” is a category to look at.
Sourcing & Selling Product Outperforms
Bringing this all together, “Sourcing & Selling Products” has the highest variance in app installs to the number of apps on the store with 7.6%. The average installs per category is more than double the app store average and it has the third highest median installs out of all the categories.
Finally, we can also be confident in this data. The table above shows that we are only not counting the installs of 13.5% of apps in the category (or 48 apps). So we can be pretty confident in the data here.
3. The drop-shipping subcategory drives Sourcing & Selling products category install performance 🚢
I hadn’t intended to get into sub-categories at this stage as there’s a lot of data to dive into here which may change slightly. But when writing this up I couldn’t leave the “Sourcing & Selling Products” insights there.
The table below shows the “Sourcing and Selling Products” category broken down into it’s sub-categories.
There are two things to note here:
“Selling Methods” has an extremely high average installs. Over 3x the Shopify store average. Whilst this could be seen as a positive, the low median installs suggests to me there’s one or two apps which have all of those installs and are driving the average up. I haven’t confirmed that by looking into the category but I’m pretty confident that’s the case. Therefore, I’m not really that interested in it.
2. “Dropshipping” however is interesting. The average installs is significantly above the Shopify app store at 3,186. But more importantly the median installs is 258. The other two subcategories have low median installs of 86, 109 meaning that “Dropshipping” is driving the high median install rate in the “Sourcing and Selling Products” category. Consequently, this is an area worth looking at closely.
Furthermore, we have c.90% of the data for the Drop-shipping subcategory so we can be really confident in those insights. (see the table below).
”Gifts” subcategory - Perfect for smaller acquisitions
As mentioned, I was hesitant to share the full category and subcategory table of installs at this point. It’s a very big table so it takes up too much room to get into a newsletter post neatly. However, if you would like the data it’s pretty easy to construct yourself in Airtable or Excel using the datasets we make available on Marketplace apps.
That said, there is one more sub-category that I want to highlight. “Gifts” in the “Merchandising” category.
At first glance, an average installs of 300 is poor given the app store average of 1,900+. However, when combined with a median installs of 265 I think it becomes interesting.
There are also only 15 apps in total in the “Gifts” subcategory so we have most of the data collected and again can be confident in the results.
This means a very high percentage of the apps in this subcategory have a high installation rate. Furthermore, there are likely no clear category winners, otherwise the average install rate would be in greater excess of the median. These are perfect conditions for smaller acquisition opportunities (less than $10k).
4. Installs & review count are positively correlated but installs & average rating are negatively correlated
One of the questions I wanted to answer is whether the number of reviews and installs are positively correlated. I did the analysis looking at correlation and the R squared number. And then I realised this is actually a pretty stupid thing to look at.
That’s because we only started collecting install data from early-mid July 2021 and installs are dynamic. Merchants continuously install and uninstall apps fluctuating the install number. Reviews on the other hand are cumulative and change infrequently. When they do change they only go up, never down.
This means an app could of had 100s of installs x time period ago and also 100s of reviews. However, today that same app may only have a 50 installs but it still has the same number of installs as at the peak.
With all that said, I decided to include the analysis in the post as I think it still tells us something albeit not the complete picture. In time we’ll have enough data to track installs with the number of reviews directly and we can revisit this topic.
What we found is that installs and the number of reviews are positively correlated with a correlation of 0.44.
Installs and average rating
In previous posts I have mentioned that the average rating and number of reviews are positively correlated but only very slightly. Therefore, I wanted to investigate the relationship between installs and average rating.
There is a very slight negative correlation between average review and installs of -0.0202.
Perhaps this is to be expected. As an app receives more installs, it is more likely to have events that lead to a negative reviews being left; slower customer service, edge cases that are harder to resolve etc.
Maybe it also gives hope to app owners that a negative review does not necessarily mean that your app won’t still get installed.
That’s all we have time for. I hope you enjoyed it and make sure you share and subscribe to get more insights