Having picked up a trick or two from mentor, former boss and The Cambridge Phenomenon co-author, Charles Cotton, the co-founders of newly incorporated startup, SaleOrigin are aiming to put them to good use in building an end-to-end analytics platform for developers of mobile apps.
During their careers, Jonathan Custance and James Green have stopped at all stations on the startup journey, both as employees and founders. They are preparing to set off yet again having stumbled across another big idea - a one-stop resource that gives developers a much clearer picture of the route taken by downloaders of their apps. In a crowded marketplace, Custance and Green figure that devs may be willing to pay a premium to identify which marketing channels are most effective.
Custance and Green share similar backgrounds, both MA Computer Science graduates from Cambridge University and seasoned tech entrepreneurs. They started their careers as BBC Micro developers before joining the startup ATML (Advanced Telecommunication Modules) in 1995, probably best remembered as Virata as it was renamed in 1998.
It was at Virata that they met mentor Charles Cotton, the then CEO who grew the company from a startup into a NASDAQ listed business, with a market cap that peaked at over $5 billion. With offices in about half a dozen countries and as one of the first companies to break the billion dollar barrier, Virata was viewed as a major contributor to the Cambridge Phenomenon.
Following a colourful history of mergers and acquisitions it eventually became Conexant Systems. But due to downturns in the telecoms market, headcount reduction began in 2002 and the UK office was closed in 2005. Custance and Green decided at this point to form their own company and created Camrivox, offering technology for simple and effective telephony integration with CRM (Customer Relationship Management) systems.
Under the guidance of Cotton, they received useful advice on speaking to venture capitalists and despite having no prior business experience managed to raise $3.5m dollars and grew the company into a 10-strong team.
Unfortunately, just as two potential buyers were undergoing due diligence, the financial crisis of 2008 extinguished the deal and the company eventually went into administration. However, Camrivox was eventually sold and its products are still available on the website.
Off the back of consulting work they did for the buyers, Custance and Green created Green Custard in 2010 (see what they did?), a consultancy that has to date developed eight mobile apps as well as offering services such as patent advice and writing back-ends for websites.
Shortly after Green Custard was formed they decided to write an iPhone/iPad app called Extreme Sheepdog Trials for fun - a pointless but amusing game that you play by whistling. It was when the time came to start promoting and tracking its success that the founders were frustrated by the lack of tools for campaign attribution, in other words, knowing where a user who arrives at your app store landing page and subsequently downloads and installs your app came from, so that you can more effectively and efficiently spend your marketing budget. They were also struck by the sheer number of various software and outlets they had to use to get an overview of its download and usage statistics.
Developed by a publisher for the publisher (meaning developers, brands or agencies), SaleOrigin – or origin-of-sales – was created primarily to address the specific need to track campaign conversion attribution. On top of that it aims to dramatically simplify the entire monitoring process by housing all user engagement analytics and app store data in one convenient location.
As the number of mobile apps continues to grow – the two dominant platforms being Android and iOS, where Android has just over 450,000 on Google Play and iOS has just over 500,000 on iTunes – the more difficult it becomes for publishers to push out a successful product. SaleOrigin decided it was therefore vital to have the tools and know-how to market and manage user needs.
In order to appreciate what all of the above means is it helpful to employ a little role-playing. Imagine that you have a ready-made app and are charged with getting it out there to as many users as possible and ultimately to make as much profit as you can from it.
For mobile apps, revenue is usually generated in any of the following ways: As a paid app at a fixed cost usually through an app store; through in-app purchases, this means allowing users to purchase additional features or upgrades directly through the app once it has been installed; or by placing advertisements within the app and earning money each time a user clicks on one.
But to get to this stage, unless you are extremely lucky and the app goes viral organically, you must employ an effective marketing campaign. You may choose to run several promotions, for example placing links to your app store landing page through Twitter, LinkedIn, placing ads through ad networks, Facebook, online articles or perhaps leaflets with QR codes – a method adopted by Green Custard’s client, Medical Apps, which sent out leaflets with QR codes directly to students.
Campaign conversion attribution is something that even experienced publishers may not be aware of – they may know how many users click on a link and total download numbers but not the actual conversion rates since app stores are effectively, as Custance puts it “black holes.”
SaleOrigin breaks down the above needs into three broad categories: Optimising campaign tracking, app store data and user engagement analytics, as can be seen in the diagram to the right.
It pulls in daily app store chart positions and measures metrics such as reviews, ratings, app price and version, and featured app listings. This allows publishers to track how these measures directly impact chart positions and enables them to easily address negative reviews or ratings.
To obtain user engagement and campaign tracking analytics, publishers need to include an SDK in their apps. Custance explains this is a simple fifteen minute job (not including update time) and requires minimal additional disk space (note that the SDK is currently available for iOS 3.0 or Android 2.1 and above. There is a Windows Mobile 7 version available but you must contact them directly for it).
More engaged users can lead to increased ad revenues or in-app purchases. SaleOrigin anonymously tracks user session data providing insight into new and returning users as well as engagement statistics. By knowing what works, publishers know where best to put ads and what additional features would encourage users to purchase upgrades.
Most of the above features are fairly bog standard. The clever part of its application, and indeed its secret, is in the implementation of tracking campaign conversion attribution. Via a user interface to its application, users generate shortened URLs by supplying parameters specific to the target for the link, which could be a blog, social media or an ad network. When an app is installed, an internally generated identifier from the device is sent back to SaleOrigin’s servers allowing it to uniquely track where a click-through for an installation came from.
It is important to note that data obtained from the device is anonymised. It can therefore track typical user behaviour but not link it back to an individual. A drawback of this is that demographic data is not provided but according to Custance, existing software is mostly based on machine learning rather than actual raw data.
Other useful features include monitoring cross-selling performance - that is tracking how many people hit a page in your app that contains links to other apps you may have and knowing how many installed. Another is referrals – if you have Facebook or Twitter integration into your app and you want to allow people to share it via a link, you can generate a shortened URL to track how effectively referrals within an app are driving downloads.
Now, device identifiers have recently become a subject of controversy. iOS devices each have what is called a Unique Device Identifier (UDID), basically a forty character string unique to each device. Due to privacy concerns, Apple has announced it is likely to deprecate its usage and apps using UDIDs will not be able to publish to iTunes. This is still subject to change but it has highlighted the need for publishers to find a suitable alternative.
There are a number of alternative methods including device fingerprinting, MAC addresses, OpenUDIDs, copy/paste bin methods (see here for an overview). SaleOrigin have not disclosed which of these it uses to generate unique IDs but has confirmed that it does not use UDIDs thereby providing an alternative to publishers concerned about this issue.
Is any of this new or are there alternatives out there? Based upon a market survey the company conducted, there are various tools that provide solutions to parts of the problem and include more detailed ad revenue data but none that do everything.
“There are several categories of companies. AppAnnie and Distimo are examples that provide a better interface to the data that comes from the app stores. There are companies like Flurry or straight Google Analytics or Localytics who provide application analytics,” Custance said.
There is also MobileAppTracking, which focuses purely on campaign tracking. In fact, with a number of players covering some of the same ground, we thought it might be a good idea to get some feedback from some Cambridge developers.
Emmanuel Carraud, CEO of MagicSolver had this to say: “At MagicSolver, we are using several third party analytics solutions, including Flurry, AppsFigures, AppAnnie, but we also built our own in-house solution to track download conversions for our app FreeAppMagic which has over one million daily active users. We work on a regular basis with over 200 iOS developers and digital agencies across the globe, so we developed our own API to track downloads through UDIDs, mac addresses, OpenUDIDs and other solutions.”
Carraud said MagicSolver is not yet considering replacing its existing systems but his experience confirms to some degree the complexity involved in obtaining a good overview of an app’s performance.
Benjamin Wirtz, founder and CEO of HandyElephant, who has recently launched a new social inbox application Unifyo said: “There is definitely a need – I struggled getting a good overview because web and app markets (and then in-app analytics) are disconnected”. Though not a user himself, he also said “I believe the vision is achievable and in contrast to other providers, SaleOrigin seem to focus on actionable analytics”.
Bango – which offers a number of tools for mobile payment solutions, analytics and campaign tracking – does not as of yet offer campaign conversion metrics.
Another app developer did not even believe it was possible to track link-to-installation conversions - a good sign if you are able to achieve the impossible as far as potential customers are concerned. Interested publishers can find examples of current clients from the SaleOrigin website.
SaleOrigin is still only a two month old company meaning the software is lacking in certain features. However it does have an extensive roadmap. In the next few months additional data sources will be added – for example, Custance believes it would be interesting to track how a link from Twitter is shared, how it progresses once shared as well as who the prominent sharers are.
While they can pull in ad revenue from AdMob, it is not currently available from the SDK. Integration with other ad networks is in the pipeline.
For the time being SaleOrigin is a free application. Though it has not confirmed the business model, it is likely to charge for the campaign tracking portion of it. The plan is to make money from those that are saving it by tracking campaigns more effectively.
For now, feedback from users is right at the top of the company’s list of priorities, so if you are a developer and would like to give the system a test drive, head on over to its website.blog comments powered by Disqus