November 19th, 2018
Just over two months ago, Branch announced our acquisition of TUNE’s Attribution Analytics. Since then, we’ve heard many questions (both from customers and others in the industry) about how the integration will work. Our team has been working hard on this project, and we’d like to share a few key details of our approach — and our progress so far.
Throughout this process, we’ve kept our initial commitment to our users to heart: Attribution Analytics customers will not need to make technical implementation changes to their TUNE setup.
We’ve made sure every user will be given the appropriate support through a system we’ve outlined below that ensures a seamless transition. This plan was built to ensure customers’ TUNE SDK and server-to-server integrations will continue to work, and the data they receive from the TUNE postback system and export API will be fully backwards-compatible, both during the integration process and afterward.
Despite this commitment, we’ve continued to hear some doubts. How can we promise such a transparent solution? Will customers be forced to implement the Branch SDK once the integration is complete?
This post answers these questions and explains the entire process in detail, but here’s the summary: our team has built behind-the-scenes infrastructure to seamlessly route all Attribution Analytics traffic via DNS to the Branch platform, without customers having to make changes.
In fact, we have already rolled out this infrastructure for some of TUNE’s largest and most recognizable customers. For the last month, these customers have been receiving live attribution data (including via Facebook MMP) directly from the Branch platform, without needing to implement the Branch SDK or make any changes whatsoever to their pre-existing TUNE implementations.
Our team still has much left to do before we can say the platform integration is complete, but we want to share a few details about our plan, and how we’re working hard every day to deliver on our commitment to a seamless customer experience throughout the process.
When our team began this project, we knew two things going in:
We decided early on to use Branch’s infrastructure as the foundation for the combined platform. This provides some major benefits:
One of the core Branch values is Customers First. That’s why before we began this transition, we made an upfront commitment: we would not require technical implementation changes by Attribution Analytics customers. The only kind of required update we consider acceptable is minor configuration changes. For example: postback IP whitelisting, which could be necessary to meet a customer’s own internal security policies.
Fortunately, the TUNE Attribution Analytics team joined Branch as part of the acquisition. This means our team now includes their years of industry knowledge and lessons learned. We have incorporated this experience throughout the integration process, building a combined product that will be far better than either of its ancestors.
During scoping process for this integration project, we identified two broad categories of work:
Attribution Analytics is a mature product and offers a lot of powerful functionality. However, most customers do not use every feature, and we knew some features would be available via the Branch platform earlier than others. At the same time, we wanted customers to experience the power of the combined platform as quickly as possible.
To meet our commitment for seamless customer experience, we decided to migrate each customer as soon as all the Attribution Analytics features they rely on are fully replicated in the Branch platform.
We began by taking an inventory of the Attribution Analytics features used by each customer. Using that inventory, we loosely grouped customers based on the features they use and the complexity of their specific implementation. Internally, we call these migration groups ‘tranches’.
Our next step was to estimate the dates by which our engineering teams expect to have each Attribution Analytics feature fully replicated within the Branch platform.
Working backwards, this allowed us to create a preliminary timeline for the entire migration project. We expect to have all Attribution Analytics customers fully transitioned to the Branch platform by the end of March, 2019.
Once the Branch platform is ready to accommodate the needs of each Attribution Analytics customer, what is involved in the migration itself?
Let’s start with the basic data flow of an Attribution Analytics customer, before migration begins: traffic comes in from the TUNE SDK (or server-to-server integration), events are processed by Attribution Analytics, and then data is available via postbacks and the TUNE export API.
To migrate this flow to Branch, the first migration step is full data mirroring. This means we begin forwarding all traffic from the customer’s existing TUNE SDK or server-to-server into our new, TUNE-compatible Branch API endpoints. This allows us to process all events in both Attribution Analytics and Branch simultaneously.
During this period, our team works with each customer to validate configurations, investigate any data discrepancies between the two systems, and help transition internal business processes from Attribution Analytics to the Branch dashboard. We’ve built internal, automated diagnostic tools to identify and resolve issues so that we all be confident that the mirrored data is processed correctly by Branch.
Once each customer’s data mirroring period is over and validation complete, our team makes a DNS update so that traffic from the customer’s existing TUNE SDK or server-to-server integration automatically flows into the new, TUNE-compatible Branch API endpoints. We stop processing events in the legacy Attribution Analytics system, and Branch becomes the customer’s new system of record.
It’s worth repeating one more time: all of these changes happen behind the scenes, with no impact on the customer’s existing TUNE SDK or server-to-server implementation.
The final migration step relates to data exports and webhook postbacks. The infrastructure we’ve built provides backwards-compatible export and postback data, which means that the customer’s existing downstream processes will continue to work without code changes. Customers will only need to swap the URL endpoint used for pulling from export API, and whitelist the Branch IPs used to to deliver postbacks (if postback whitelisting is part of their Security policies).
The process above is not just theoretical: Our team has completed end-to-end testing of the entire system, including successful migration of internal apps. Some TUNE customers are already using Branch to report on campaigns links and attribution setting originally defined in Attribution Analytics, while running Facebook campaigns defined in the Branch dashboard! That’s right, Facebook campaigns have returned.
We’re excited about rolling out the power of the Branch platform to even more users – and we’ll be sharing more updates about our progress soon.
Ready to find out how Branch can help improve your attribution strategy? Contact our sales team, and we’re happy to tell you more.