.webp)
The following is a two-part blog on aiding your decision between building and buying a B2B marketing analytics and attribution solution. Part I deals with understanding the need for B2B marketing analytics and attribution. We also breakdown the technical requirements for an in-house solution, and what it would take for an off-the-shelf solution to deliver a similar experience.
Part II takes a logistical standpoint. There, we explain the practical resources required for a robust B2B marketing analytics and attribution solution. We also break down the opportunity cost of building and buying such a solution. Hopefully, this leaves you with valuable insights in cementing your build vs buy decision.
1. DEFINING THE NEED
Before we get into the nitty-gritty, we need to understand the need for marketing analytics and attribution under a business’s marketing function. This in turn necessitates the requirement of an off-the-shelf or in-house solution. This can be boiled down to two needs — tracking and optimization.
Tracking: In recent years, marketing has been losing its gravity as a core function of a business’s operations. As a result, justifying its revenue contribution is becoming increasingly vital. With lengthy B2B sales cycles that stretch across several months, waiting on revenue is simply not an option. This is where the use of marketing analytics with indicative metrics and KPIs (such as CPA, CTR, CPL, and web traffic analytics, etc) help in tracking and justifying the ROI of the marketing function before having to wait on your closed-won revenue contribution.
Optimisation: The optimisation need can be summarized as the need to quantify performance at a channel/campaign level and determine what to invest more or less in. Multi-touch attribution is used to facilitate this need. The requirement to optimize B2B marketing efforts will vary depending on the length of your sales cycle — but is nonetheless indispensable. Attribution is a tentative requirement, while Tracking is usually an everyday thing.
2. UNDERSTANDING YOUR REQUIREMENTS
Now that we have defined the need for a marketing analytics and attribution solution, choosing to build or buy comes next. The problem here is acknowledging that when you’re considering a solution, different marketers have different needs and resources. While some marketers only need the bare minimum, others may opt for more sophisticated solutions.
The presupposition here is that the following part of the blog on your technical requirements will always highlight more than just the bare minimum, while also zeroing on what an off-the-shelf alternative has to offer.
TECHNICAL REQUIREMENTS
Tracking and Collection
Your primary requirement for an in-house marketing analytics and attribution solution is to track user data and user interactions. There is a wide array of data that is to be tracked — page views, URL changes, web events, web sessions, CTAs, button interactions, form downloads, demo scheduled, and more. While you could limit what you track, it’s always advisable to track as much as you possibly can, as long as it is relevant to your analytics.
Ideally, your in-house solution should be able to track all of the above. There are two ways of building this. You could either opt for a CDP like Segment to automatically collect this data or construct a solution with developer dependency.
The developer will need to be able to create an event into your analytics stack every time an action occurs. The biggest concern with this approach is that it is risky. Especially when there are changes on the website, as it is too meticulous to be able to scale. That being said, you want to have a solution where your marketing team is not too dependent on your engineering team. Else this will only cause more harm than good.
SDK Requirements
The general concern with using an SDK is understanding that it will have to work with different website frameworks — like single-page applications (SPC) and normal web applications. Additionally, as websites are developed using different technologies like React JS, WordPress, CMS HubSpot, you’ll have to ensure that your SDK is sufficiently compatible.
Your SDK will need to be fine-tuned to be able to send data to your server. Webpages have to load A LOT of content including images, animations, text, etc. Your SDK will need to be able to capture events and put them in a queue to be sent at a later time. It will also have to be optimized for different internet speeds — mobile internet vs broadband. These factors should not be taken for granted. Failing to optimize your SDK could either crash the client website or result in the collection of incomplete data. Your goal here is to engineer a light yet effective SDK that captures and sends data without a compromise on user load time.
Tools like Segment and Google Analytics will help you track the events. However, these solutions are not fully automated and will require developer dependency. Factors.ai is tailor-made for marketing analytics. We also facilitate an advanced SDK that empowers robust web tracking, zero developer dependency, and more.
Data Handling and Cleansing
For a marketing analytics use-case, your in-house solution will need to be able to handle different types of data. More specifically 3 types of data:
1. Campaign reports from ad platforms
2. Event tracking data from your website
3. Objects from your CRM
Your goal is to build a solution that can handle all 3 of these data types. Doing this will prove to be challenging. At a base level, you will need to understand the objects of your CRM and how they are connected. Salesforce, for example, records accounts, contacts, opportunities, leads, products, campaigns, users, and dashboards as their standard objects. You can also create custom objects. You will then need to model your web analytics events against the users on your CRM customer data. Not to mention a separate data stack for ads data from Facebook, LinkedIn, Google Ads etc.
Data cleansing could be best explained with an example: In most off-the-shelf web analytics solutions, a web session will have some parameters set to distinguish itself from other sessions. For example, when a user is inactive for more than 30 minutes on your website, it will be tracked as a new session. Another one is when a user visits your website through an ad and a session begins, but then clicks on another ad to the same site shortly after. In this case, they will be considered as separate sessions as they come from different ad sources.
In other words, designing web sessions based on a period of inactivity or distinct UTM parameters are examples of data cleansing. Failing to do this, and other such data cleansing practices, will result in a lot of nuanced difficulties.
While most off-the-shelf solutions will handle such cleansing and data categorization, note that most of these analytical solutions usually handle only one of the three data types. Factors.ai, on the other hand, consolidates all three.
Data Storage and Warehousing
As a business running data analytics, it is important to acknowledge that data warehousing is a core need. Unfortunately, companies that adopt a data warehousing solution are still in the minority. One could argue that they could rely on application databases. This, however, will result in processing constraints and logistical difficulties. Alternatively, most data warehousing solutions process analytical queries in a more effective, columnar fashion. They also serve as a centralized data hub for all your workflow data. Modern data warehouses also make it cost-effective to scale your data warehousing. Therefore, it would be preferable if your analytics solution had an export to a data warehouse like Google BigQuery.
Attribution and Presentation
The previous tracking and handling data requirements form the basis for a robust marketing and web analytics. But what about marketing attribution?
For an attribution solution, you will first have to refer to the contacts on your opportunity account in your CRM. Then identify all data touch-points you encountered with those contacts — these could be webinars attended, demos scheduled, white paper downloads, field events, etc. These are all website sessions that are driven from different ad campaigns, email campaigns, etc. After accumulating all this data over a certain time frame, there will be an X amount of opportunity value that can be attributed to these touch-points. You will have to credit them and to do that you can utilize several existing multi-touch attribution models — refer to this blog to learn more. You could even implement a custom model — for example a model that attributes more credits to contacts with job titles past a certain level, and fewer credits for one below that level.
Presenting your marketing analytics and attribution reports requires breaking down, summarizing, and visualizing an extensive amount of data. For this, you would have to set up a dashboard and operate a data visualization tool — like Tableau and Looker. This requires a fair amount of expertise to assemble. What makes this process challenging is building out all the SQL queries for these reports.
Ensuring the correct data and the right quantity of data is being delivered to your dashboard is key. An overcomplicated dashboard or several superfluous dashboards won’t run efficiently. Ideally, you don’t want your team of data engineers to be preoccupied with operating existing data pipelines.
Factors.ai comes with a powerful attribution engine. With it, you can use, compare, and customize several single-touch and multi-touch attribution models. Factors attributes touch points across ad platforms, website events, and CRM. Factors also supports real-time reports and insights, a unified and customizable dashboard, and a wide range of data visualization under one roof.
Maintenance
When you build an in-house marketing analytics and attribution solution, there is a need for continual maintenance to ensure operational efficiency. Especially the maintenance of integrations across your SDK, CRM, Ads platforms, API version updates, etc.
Optimizing your SDK from an engineering perspective as mentioned before is a trial-and-error process. You will have to adjust your data pipeline to effectively deliver data to your dashboards. You’ll even have to keep your techstack up to date. Not to mention that there will always be tech debt and bugs to troubleshoot over time. This is never a one-and-done situation. As time passes you would have to modify your queries while you bring in more data all while optimizing your process.
Integrations, data concerns, troubleshooting and all of the aforementioned maintenance can be administered without developer dependency with Factors.ai.
3. INTERPRETATION
The purpose of part I is to illustrate the challenges in setting up the technical requirements for building a solution — optimizing your SDK, building a data layer and a solution that is compatible with all types of data, building a solution that can export to a data warehouse, assembling a dashboard, maintenance, etc. This way you can have a fair idea of not only what to decide on, build or buy. But also know what to look for when opting for an off-the-shelf solution.
If you were looking for a comprehensive conclusion, then look no further than part two. The next part highlights the opportunity costs for an in-house and off-the-shelf solution and takes into account the cost, time, planning, and even the technical requirements of this piece.
