Published: 9/27/2022

“Code is a liability. Every line of code is a commitment: to maintain it, to explain it; to a direction.” -Kevlin Henney.

Marketing Intelligence Data Science and AI at Scale

There is no shortage of open source and commercial tools in data science, machine learning, and AI. Inevitably, this spurs the age-old software question of build versus buy. In the realm of marketing intelligence and analytics, building internal tools comes with a few caveats.

The idea of a standalone, pre-built solution that will work consistently for marketing data is becoming outmoded. Beyond the investments of time and resources required to build an internal tool, which is considerable, consider the following:

• You must have extensive domain expertise and familiarity with the marketing data available from Social Media and Web analytics companies.
• Requires a unique range of skills to custom build, maintain, and improve in-house marketing intelligence and analytics tools. For example, each marketing channel, such as paid search, content marketing, and email marketing, introduces additional API complexity, new data types and providers, and specific business domain knowledge.
• The proliferation of marketing data providers and technologies (MarTech) has exploded. Combining Social Media, Web Analytics vendors, OTT Advertising (Hulu, Roku, Paramount+), Podcast streams, E-commerce data, and transactional data with unique APIs is complicated.
• Adding to this complexity are the unique data structures and metrics that need to be extracted, transformed, and loaded per data source basis. The raw data stream in a normalized form must exist before analysis can even begin.
• Ongoing maintenance of API integrations with third-party Data Providers (social and web analytics companies) requires dedicated staff to maintain any changes made by the source providers.
• In 2011 there were approximately 150 data-producing tools, and by 2017 there were over 5,000. According to Full Circle Insights, over 8,000 MarTech solutions are available worldwide. Choosing a vendor with the appropriate solution can be difficult.
• Third-party cookies are eventually being phased out due o adopting a solution, like Alembic, which does not require additional cookies, preserves customer privacy, and approaches Marketing Intelligence in a revolutionary way.
• Rate limiting and continued access are significant factors in data collection. Alembic’s unparalleled knowledge of social media APIs and high-profile customers ensures that Alembic can continually import data for your accounts with a lower chance of being blocked or rate limited during these operations.
• Time series reconstruction: Nearly every vendor offers a “lifetime” total of essential metrics such as impressions, engagements, and reach. Alembic is the only vendor capable of reconstructing the hype and flow of a post’s impact.

Alembic is an enterprise-grade SaaS Marketing Intelligence platform that employs a unique approach to addressing the caveats above. Find out if our product is right for you.

The Burden of Building an In-house Marking platform.

Developing an in-house Marketing Intelligence Platform requires significant investment in resources. In general, an organization will need to invest in:

Where do you store all your data?

Storage and processing are at the bottom of the data/AI “hierarchy of needs” – see Monica Rogati’s famous blog post here – all of which must be designed, built, and tested before you can begin to apply data science techniques. Therefore, one must consider the significant investment in creating and maintaining an AI/Data Analytics technology stack dedicated solely to Marketing Intelligence when considering the factors that make up the build vs. buy decisions.

The short version of the stack includes:

1. Extract/Collect: Third-party API support, external data ingestion, and user-generated content.
2. Transform: Clean the data, normalize the data, detect anomalies, and prep the data to be loaded.
3. Reliably load and store: Create a reliable data flow, stable/scalable infrastructure, with monitored pipelines that will support structured and unstructured data.
4. Aggregate/Label: Understand the metrics and how they will support the analytics, aggregates, features, and training data.
5. Learn/Optimize/Explore A/B Testing, experimentation, simple ML algorithms, and statistics.
6. Data Science: Deep learning and insights throughout the lifetime of the dataset

Given that resources are limited, it is essential to ask: Will an in-house analytics project be your organization’s next prominent growth provider?

It’s essential to consider what you’re building and which specific resources are required. For example, it’s not likely that your developers will create a database but rather license an SQL-type, commercially-available cloud solution. For visualizations, your developers can use a library like D3, or you can authorize a BI tool, which you will also need to maintain.

Regardless of what libraries are used and which software is licensed, combining internal tools will build tribal knowledge. The staff will need to maintain their skills with the technology stack selected. Only 23% of respondents prioritized spending time on skill development, the lo west rank among all other analytics activities. The most devastating problem here is that if a critical developer leaves your company or is transferred to another department, you could be left with a non-working tool and lost capital investment. Alembic normalizes all the event streams connected to the platform, recreates time-series data, detects critical changes in the aggregated data, and then correlates the detections to provide fundamental insights based on data science maintained and pioneered by Alembic.

If you’re considering building in-house for all or part of your marketing analytics solution, keep in mind the team or individual that produces it. They will need to:

• Maintain the software costs and updates for all software and libraries used to create the solution.
• Maintain subject matter experts, product experts, software developers, QA teams, and Product Management for the Product Life Cycle.
• Be available for project-like updates of undetermined length daily or weekly.
• Train staff in the meaning of analytics regularly.

For some, building an in-house analytics application makes sense. But in the evolving world of marketing data, it’s essential to understand the risks and requirements that go with it.

Accelerated Time to Market: Alembic

According to Gartner, 52% of marketing leaders still spend most of their analytics time on data management (preparing, formatting, and integrating). This time-consuming approach has left marketers desperate for a more straightforward, efficient, affordable, and reliably accurate solution to tie sales and marketing activity to revenue so that they can better spend their time instead of cleaning up and piecing together data.

Alembic can help you break free of the time-consuming burden of data management and direct your marketing dollars to drive insights starting on Day 1. We handle all the complexities so that you can focus on the senses. What’s more, we can have your Social Media and Google Analytics accounts linked to Alembic in under 30 minutes without involving your internal IT department. Once your account is linked to Alembic, the ETL process will commence. Typically, your Alembic account will be populated with your data within an hour and derive increasingly meaningful insights from 48 hours onward. Each data stream connected will provide insights at varying times from ingestion. Alembic will walk you through how to link your social media and web analytics accounts to Alembic, allowing you to generate actionable marketing intelligence for your brands and company.

• Automatically normalize all API-linked event streams.
• Recreate your data in time series to detect outliers in the aggregated data.
• Time series reveals content resurgence, stickiness, and engagement over time.
• Correlate event detections allow for the discovery of actionable insights.
• Understand what’s changed, what’s working, and what’s driving conversions to more accurately direct marketing spending to drive ROI.

Purchasing Alembic removes the roadblocks by:

• We provide lower upfront costs than building and staffing a complex in-house data science platform.
• Ensures stakeholders gain insights into marketing efforts in weeks rather than months or years of requirement wrangling, budgeting, tool selection, staffing, and software development typically associated with the software development lifecycle.
• We are reducing ownership costs by eliminating the need for your staff to maintain domain expertise at the API level.
• We are providing a higher ASP through added capabilities. Alembic’s sole focus is on developing the most sophisticated Marketing Intelligence platform.

What does this mean for marketing leaders? Marketing is responsible for critical customer-facing, revenue-generating systems and applications. It means that, if they haven’t already, they ought to appoint a chief marketing technologist in title or role equivalent to look after this growing technology estate. And, crucially, it means marketing is responsible for collaborating with enterprise IT, ensuring coordinated, efficient, and harmonized technology spending. Or they could subscribe to Alembic and be up and running today.

Conclusion

Ultimately, the decision to build an in-house analytics system will come down to your assessment of the pros, cons, and your own goals.

Building an in-house data analytics platform offers the promise of customized dashboards, control over functionality, and unique data integration at very steep costs in terms of resources, expertise, and time. Building analytics is code intensive, which means more developer resources are committed not only to the development but to the maintenance of both the in-house code and third-party APIs.

The buy solution – Alembic offers:

• Faster time to a deployment where time-to-market is crucial to success
• Frees up resources so your development team can focus on other priorities
• Requires no IT or Engineering resources to onboard
• Requires minimal ongoing administration
• Provides performance and scale for all analytic offerings
• The skills and training needed to gain immediate insights are greatly simplified.
• Provides the only Event Correlation engine for marketing on the market today.

Even at the most straightforward analysis, building a cloud-based marketing analytics platform by leveraging pre-built analytics stacks will have a TCO (Total Cost of Ownership) for a medium-size business that is far less than building an in-house system in the first year; especially when factoring in the costs of staff and the time to develop and deploy.

Alembic can provide you with an on-demand marketing intelligence platform at a fraction of the cost without defocusing IT or internal data analytics teams.