The UCOVI Blog
The UCOVI Blog
Debate: Should Data Analytics teams sit within Sales/marketing or IT?
Ned Stratton: 26th October 2021
As a data analyst I see myself as the "bastard child" of an IT guy and a greedy CEO; taking a great (possibly unhealthy) level of pride in the technical side of my job, but also wanting to uncover a cash-falling-from-the-sky scoop in the numbers I'm crunching. Or at the very least see that my reports actually contribute to a decision by someone (ideally a correct one).
I'm mixed about this. Even having made the journey from "decent with Excel" to "knows what a SQL stored procedure is" to "Python/Power BI/SQL/own website that's not a Wix template", I’m still marketing's token techie, not bona fide IT.
On the one hand this is good. The people who work in marketing teams are fun, and I learn about the business I'm analysing data for more quickly (domain expertise being a vital third of the holy trinity that also includes technical knowledge and maths). But I'm cut off from reams of others who share my professional interests and have technical things to teach me. Worse, as an analyst I constantly find myself needing access to things called "databases", or new Python libraries/bits of software. Since these require computer administrator passwords to install and I'm not considered to be IT, I need actual IT to install them for me, which exposes me to the circle of hell known as "Priority 4" new software request tickets. Put simply, part of me longs to be in the IT team! DBAs and data engineers get to be there, so why can't us analysts?
What business function do data analysts work in?
Going by the companies I have worked in as a data analyst over the past 6 years, the analytics team always sits within the commercial function (ultimate reporting to the Chief Commercial Officer or Chief Marketing Officer), within the marketing team, or (in bigger companies) decentralised and split up into specific analytics teams for marketing, product, operations and finance. It has never reported into the CTO, whereas other data roles such as database admins, database developers or data engineers have.
Available stats to show if my experience reflects the broad trend are sadly and surprisingly thin, especially for a topic that an old CTO colleague told me was an "age old debate". The best I found was from an American body called the Advertising Research Foundation. They produced a 30-page "Organizational Benchmark Study" in 2020 that collected survey data from the advertising sector. It found that only 1% of data science, research and analytics teams within the sector reported to the IT team, as opposed to 7% reporting to marketing, 24% to consumer research and – schoolboy-survey-design-error drumroll please – 36% to "Other".
This finding only covers one sector, and the survey sample size was small (176 replies). So digging deeper I found that the UK recruitment firm Reed has a free job postings data API (who doesn't love a free data API?), and pulled off some 12,000 live job postings with the keyphrases "data analytics", "business intelligence" and "customer insights". I redacted this down to 800 that were actually relevant to the keyphrases and were senior-exec, managerial or department-head level, and managed to pull out 28 from this that had the director-level title to which the vacancy reported in the job description. 11 of the vacancies reported into a commercial function, whereas only 6 reported into IT, so more or less double. (The other 11 reported into a dedicated Director for data or Chief Data Officer, a fashionable board-level development that I'll cover later).
So my two pots of evidence, though small, broadly support or at least don't contradict my experience that marketing is currently holding the cucumber when it comes to analytics.
What business function do data analysts think they should work in?
This was far simpler to gather evidence for owing to the single source of workplace-gripe truth - LinkedIn polls. Some polls I ran recently would suggest that data analysts and scientists are split down the middle on whether they're commercial or IT. I asked the question "Should Data Analytics teams in companies sit within and report to a commercial function (sales/finance/marketing) or IT?" in two LinkedIn groups (Data Analyst Forum and Data Scientist & Analyst). Across both, 190 voted for commercial and 164 for IT.
But this 54% majority for commercial differs from the results of a similar poll I ran in anger a few months ago, where I asked the data analysts and scientists in my LinkedIn network whether "Business Users" or "IT/Security teams and process" were the biggest drain on their job satisfaction. 55% polled for IT/Security teams and process – something I interpret as a coded plea from analysts to be members of the IT team to have more influence over processes and people that affect their work, or be taught the short steps needed to run said processes for themselves. Part of the frustration of being a computer-literate data analyst outside of the IT team is that I have the ability and knowledge to install MySQL Workbench on my computer, or automate the emailing of daily sales reports from the company Exchange server – all of which would free my time for profit-making insights – but I just don't have the permissions to do it. To cap this, I need to wait behind computer lockouts and other named priorities to have someone else do it for me.
Where do commercial leaders think data analytics should be?
Business leaders see their own departments as the natural home of data analytics, with the justification being that the greater the proximity between data analysts and the users of the work they produce, the more insightful, relevant and user-intuitive the analytics produced will be.
Ewa Campbell, a CMO and marketing technology expert within the events and media industry believes that "to maximise business results – sales/marketing/data (and customer service) should be one connected function. Those 3 teams need to fully understand each other's roles and their impact on business results. So, if efforts are combined and processes streamlined between those 3 functions then they can drive real impact. IT is a separate story."
Data analysts – who would otherwise lose themselves in detail, pursue any and all lines of investigation, or prioritise dragging their favourite Power BI custom visual kicking and screaming into the analysis - need maximal direction from marketeers, product owners and other subject-matter experts to guide them through the data they explore, giving them the important questions to answer and assumptions to test.
Where do CTOs and technology leaders think data analytics should be?
Two schools of thought emerged from the tech and IT leads I spoke with.
Richard Bastin, a CTO in finance and David Johnston, a London-based BI consultant and developer, both supported the idea of data analytics being an arm of IT to achieve the dual benefits of efficiently upskilling data analysts from a technical perspective, as well as benefitting from their influence on the wider IT function. In Richard's words, analysts being in IT would "ensure they leverage the rest of the function and learn best practice", while David Johnston made the point that the broader tech function within a business "will generally be far more likely to consider the Analytics team when implementing new processes or software".
However, some on the tech side believe that Analytics should be its own function separate from both IT and other business divisions. Marc Firth, an Engineering VP at an online marketing firm believes that "Analytics is its own commercial function. They have their own concerns with providing insights to clients and the other functions, i.e. managing analytics properties, generating reports and dashboards, and building models that can be used by the clients and other functions to make better informed and automated decisions."
This mirrors the recent rise of the Chief Data Officer, and an even newer phenomenon of Chief Analytics Officers reporting directly to CEOs.
The UCOVI Verdict
The Chief Data Officer route seems the most superficially attractive in that it gives data the board-level attention and credibility it needs and deserves.
However, it has two drawbacks. Firstly, a CDO's priorities tend to lean towards issues around privacy, security and overall data governance, or "keeping the CEO out of jail". Analytics for revenue opportunity and business growth with take second preference to this, so data analysts won't get the board-level advocacy they'd like. Secondly, I believe the Chief Data or Chief Analytics officer positions are part of the wider Chief Anything Officer corporate trend of putting support services like HR, data analytics and IT onto the board. It sounds wonderful, but in reality it serves to dilute a senior management team, and gives the Sales, Marketing and Finance directors some whipping boys to blame their own functions' shortcomings on in front of the CEO.
To accurately place data analytics within an organisational structure, we need to be clear that it is there to support the primary business functions (and sometimes even IT and InfoSec). It then becomes a matter of ensuring the analytics team is a. correctly steered (they do work that benefits the company) and b. given the tools and structure they need to provide this support.
A data analytics team should therefore be built within the IT function of a business and report ultimately into the CTO.
This way, data analysts can bring business perspective and bigger picture thinking to the wider IT team, while simultaneously learning best practice programming methods from more technically accomplished and experienced data engineers. On the inside, they can influence core business processes that collect data to make sure the data collected is rich enough to produce insights, instead of enriching and wrangling the data in external, crash-coded analytics ETL workflows that are often undocumented within IT and lead to technical debt and duplicate processes.
Stack Overflow's excellent 2021 Developer Survey (raw response data available for download) provides a hidden gem of evidence to support this. It collected responses from 80,000 Stack Overflow users (developers, computer science students, data engineers and analysts) on numerous questions including the coding languages, databases, and technologies they most use and want to use, but more interestingly what they do when they get stuck.
Redacting the responses down to the 49,000 who were in any full-time software, engineering, IT or data role, I split them up into the 2,745 who said they were a "Data or Business Analyst" and the other 46,000 who didn't. When answering the multi-tick box "What do you do when you get stuck?" question, the Data Analysts were more likely (by about 6%) to opt for a self-help solution such as Stack Overflow articles or tutorial videos than other technical role performers, and nearly 10% less likely to opt for a collaborative solution or mentorship from a peer. This shows that data analysts, isolated from their technical peers, struggle in silence and cut and paste solutions that they don't always understand fully. This entrenches bad practice early on in a career.
| Stack Overflow Survey 2021: Getting stuck | ||
| Data Analysts | Other full-time roles | |
| Visit Stack Overflow | 84% | 79% |
| Watch help / tutorial videos | 46% | 38% |
| Call a coworker or friend | 41% | 49% |
As for the steering influence of business users on data analysts that comes from being in sales and marketing? Data analysts can motivate themselves to be hugely inventive and technically more advanced than their last task from within a commercial team. What's to say they can't do the reverse and find inner reserves of business savvy from within an IT department? Furthermore, once the CTO is accountable to the board for insights from data as well as keeping the lights on, the imperative will cascade down from them to ensure the output of their analytics team is useful.
There is even evidence that a data analytics team can be steered negatively by virtue of its presence within a commercial function. In 2020, Gartner polled more than 400 marketing leaders and analysts to produce its Marketing Data and Analytics Survey (downloadable here), which reported on page 10 that the joint top two reasons for Analytics not being used for decision making were "Poor Data Quality" and "Data Findings Conflicting With Intended Course of Action" (both 32% of responses). To any analyst that's been death-glared when their report gave the wrong answer, this will ring a bell.
Within IT, a crack team of data analysts is better placed to both improve data quality and produce the editorially independent insights from data that a business needs to turn its fortunes around.
⌚ Back to Latest Post