Marketing Data Analysis Tips: 51 Marketing Data Analysis Tips and Tricks, Analysis Techniques, and More

Marketers for companies of all sizes rely on marketing data analysis to drive decision-making, forecast outcomes, evaluate the effectiveness of marketing campaigns, identify market opportunities and potential new audiences, and much more. But marketing data analysis can easily be overwhelming, and not only because of the massive volume of data that most companies have to work with but also because there are seemingly endless tools, techniques, methodologies, and best practices that can be applied to marketing data analysis.

To help you sift through the noise, we’ve rounded up 51 tips from leading data analysts and marketing thought leaders. These tips will help you refine your focus, identify the metrics that matter, apply the right analysis techniques, and present your findings in a meaningful way and take action that generates results.

Use the links below to jump to tips in a specific section:

Marketing Data Analysis Tips and Tricks

  1. Use data discovery in the planning process. “Data discovery is the process of identifying the data sets and data streams that can become a part of your big data initiative. Most often, organizations possess numerous data silos. Data silos are systems in which data is stored, locked away, in a sense, from other databases within the organization. Data silos can be anything from the spreadsheets used by finance to the legacy system used by production to website platforms and the marketing systems listed before.

“Data can also be found in your mobile apps (internal and customer-facing), the text documents and pdf files around the office, and all of your previous marketing materials and metrics. Any data in any form that could be useful to marketing should be included in data discovery.” Breaking Down Big Data: 4 Tips to Simplify Data Analysis, ReachForce; Twitter: @ReachForce

  1. Start with the basics and grow from there. “You don’t have to delve deep into data science to do more successful marketing. Marketers who are new to data can start with website performance data generated by free tools like Google Analytics and Google Webmaster. How many people are converting on your landing pages? How many people are clicking on your Facebook ads? What percentage of your target audience open your emails? Get comfortable with using numbers to demonstrate results!” – Ying Yi Wan, Marketers! Here Are the Must-Know 2017 Digital Trends: Part 2, Construct Digital; Twitter: @constructsg
  2. Assess the quality and accuracy of your data. “Over half (55%) of the marketers we surveyed believe that the quality or accuracy of available data is critical to driving marketing decisions, but only 36% said they were extremely satisfied with the quality of their data. Meanwhile, early movers are already deriving actionable insights from their data and your competitors could be among them. This is low-hanging fruit in the grand scheme of things and needs to be addressed.” – Steve McNicholas, Five data marketing tips for staying ahead in 2017, Callcredit; Twitter: @Callcredit

  3. Stay focused. “Focus your research questions and what you are testing. You don’t want to ask too many questions (i.e., test too many things), or it can become difficult to analyze the data due to possible interactions between variables.” – Vann Morris, Data Analysis: B2B Marketing Tips, MLT Creative; Twitter: @MLTCreative
  4. Act on your data. “Doing the marketing data analysis and finding all the information is pointless if you decide not to act on it. You gather all the data and analyze it. You come up with solutions and improvement ideas. But if you don’t implement those ideas then it is pretty much a waste of your time and energy.

“Marketing data analysis is basically a dominoes effect. You find out the information. You act on that information. This in turn leads to improvement in performance levels. Thus, your whole company starts to get better. Marketing strategies become better and so does your management.” Reasons Why Marketing Data Analysis Is Important, PESTLE Analysis; Twitter: @PESTLEAnalysis

  1. Plan for ROI. “Value exists in quantifying the expected outcomes from marketing investments. Learn what to measure, when to measure and how to measure. In order to achieve your goals, establish specific steps to move the process along.

“According to DialogTech’s 2011 State of Marketing Measurement report, 82% of marketers say their executives desire every campaign measured, but less than a third can effectively evaluate the ROI of each channel. Moreover, only 48% of marketers are using web analytics tools to measure marketing campaign effectiveness.

“Find clarity in your plan by creating an initial outline. Look at historical data; pinpoint any trends. Then, flesh out your outline into a detailed plan. Figure out how you can install analytics into your existing process, like sending marketing emails and launching new products.

“Realize the importance of measuring ROI to receive the full benefits of your marketing dollars. ASK: How are you going to reach the anticipated ROI?” – Shayla Price, 5 Tips to Increase ROI With Marketing Analytics, Kissmetrics; Twitter: @Kissmetrics

  1. Don’t fall in love with fancy charts. They should be the exception, not the rule. “Knowing how the data is going to be presented will help your analysts avoid wasting precious computational time making fancy charts and graphs if you only need the information for internal use.

“Formatting of charts and graphs can end up taking way more time than one would imagine, so an analyst should worry about pretty charts only when needed.

“Another reason it is important to discuss how the data will be used is because your analyst might use a more efficient reporting structure. They may use graph and chart types that you ask for when in fact, they could have used a more sophisticated technique if they knew what the end reporting needed to show an audience visually.

“For instance, conversations that ask, ‘Do you need bar graphs for each individual variable?’ should happen a lot more often than they do.

“This can become cumbersome and meticulous leading up to final presentations, but if the information is represented with clarity and efficiency using the right combination of charts, everyone wins.” – Benjamin Fillip, Marketing Analytics: 4 Tips For Productive Conversations With Your Data Analyst, Marketing Experiments; Twitter: @MktgExperiments

  1. Opt for the mean over the average. “Often times we hear statistics based on the average or the mean value of a dataset. The average, however, is often not the most representative number and should be used with caution. This is because the average includes the most extreme values in a dataset, i.e. it gives equal weight to all numbers including the smallest and largest values.

“The median, however, does not.  The median is just the data point in the middle.  It is not affected by wild outliers and, because of this, it is often more representative of the data than the average.” – Melissa Miller, Why the Mean Is Killing Marketing Data Analysis [Excel Tips], HubSpot; Twitter: @mcdmiller, @HubSpot

  1. Leverage data to identify your target personas. “One of the main reasons to invest in a data-driven content marketing strategy is to gain the best possible understanding of the target audience.

“A solid content marketing strategy can bring the audience closer to the brand. This can only happen through a framework that takes into consideration the audience’s habits, preferences, and needs.

“An analysis of the available data can help marketers draft more relevant personas, which helps in tailoring content to the target audience.

“Data can provide useful answers to questions such as:

  • the customers’ reaction to the existing content
  • their favorite types of content
  • their preferred methods of communications
  • the channels they are using
  • their browsing habits

“This should be the start of an effective content marketing strategy, setting up the groundwork for a data-driven approach that relies on insights, rather than assumptions about the target audience.” – Tereza Litsa, 5 tips to create a data-driven content marketing strategy, Search Engine Watch; Twitter: @sewatch

  1. Use data from Twitter and other social channels to further define your target buyer personas. “With the right tools, Twitter data analysis can help you gain insight into your current Twitter community, which can be used to help you outline and define your target personas. Applying Twitter data analysis techniques to a Twitter community opens the additional insight into the type of people your Twitter marketing is attracting and help you further define their profiles with information such as; what they’re interested in, where they’re located, and more.” – Rachel May Quin, 6 Ways To Apply Your Twitter Data Analysis Findings, Audiense; Twitter: @AudienseCo
  2. Scrub your data prior to analysis. “To perform the analysis for our report, I spend around five hours scrubbing the data to make sure it’s ready for analysis. Scrubbing can involve removing bad data or test data. It can also involve normalizing the data that was entered in varied ways. As you gather the data, you want to analyze, plan to spend time scrubbing the data. As an example, it’s not uncommon to see United States, United States of America, USA, and U.S.A. and the same can happen with other countries. Taking the time to clean up these values will allow you to perform an accurate analysis later on.” – Mike Moore, Data Analysis Tips from the State of Inbound Channel Marketing 2016, Averetek; Twitter: @Averetek

  3. Know your cost per acquisition. “Your cost per acquisition – and this is across all your digital campaigns (non-digital too, but we are an internet marketing agency, so that’s what we’ll focus on). Everything has a CPA, not just your PPC program or other paid advertising. If you’re paying someone to manage your social media channels, that’s a cost… figure it out.” – jcormier, Web Analytics – Data Analysis Tips, WSI; Twitter: @P_G_F

  4. Conduct an industry analysis. “Whether you’re entering an industry for the first time or looking for ways to take your business to the next level, it’s important to conduct regular industry analyses. An industry analysis is meant to help you review various market and financial factors in your industry that affect your business, including evaluating your competition.

“Industry analyses are useful in a variety of fields ranging from manufacturing to retail, and involve multiple factors including geographical area, industry outlook, regulatory environment and target audience. By investigating and analyzing your competitors, you can determine the best strategies for achieving business success.

“There are two principle methods by which businesses can conduct an industry analysis. The first is a quantitative analysis, which involves the use of mathematical forecasting to assess data. The second is a qualitative analysis, which requires owners to use their own judgment when reviewing information. Once you’ve assessed your competitors’ strengths and weaknesses, you can identify and implement strategies to boost your own company’s performance.” – April Maguire, Market Research Tips: How to Conduct an Industry Analysis, QuickBooks; Twitter: @intuit

  1. Capture data from both online and offline campaigns. “Offline campaigns are the hardest marketing campaigns to track. However, they should still play a significant role in your marketing mix. If you’re running print, TV or radio ads, a useful way to track those conversions is to offer your customers a discount if they input a certain code upon registration. The incentive will entice them to type the code in and let you know where those conversions came from.

“Alternatively, you could create a custom microsite off of your domain or landing page and ask people to search or type in the URL. If your website is www.abc.com, create a microsite on www.xyz.com.” – Gil Allouche, 4 Marketing Tracking Tips for B2B, Metadata.io; Twitter: @metadataio

  1. Master the four fundamentals of growth. “The fundamentals help you solve the variety of growth problems you will face in your specific situation. The better you are that fundamentals, the harder the problems you will be able to solve. The fundamentals for growth are not Search, Facebook Ads, Content, etc. It also has nothing to do with a set of tools.

“The fundamentals for growth are:

  1. Data Analysis – Understanding the meaning of data to identify, understand, and pinpoint solutions.
  2. Quantitative Modeling –  Translating your historical data into a forward looking model to simulate the future which helps you understand better what you should be doing today.
  3. User Psychology – The data doesn’t matter unless you know how to influence the numbers in an authentic way. That comes down to understanding the psychology of your audience, problem, and solution to figure out what your users respond to.
  4. Storytelling – You can be the best at quant modeling, analyzing data, and understanding the psychology, but to be a great growth professional you need to know how to bring those things to life in a way that is compelling and interesting to your target audience. This is storytelling.

“Everything else like Copywriting, CRO, Facebook Ads, etc. build off the foundation of these four things.” – Brian Balfour, 7 Principles To Mastering Growth Marketing, BrianBalfour.com; Twitter: @bbalfour

Marketing Data Analysis Techniques You Should Know

  1. Know when (and how) to use correlations. “Correlations are used when you want to know about the relationship between two variables. For example, you want to know consumers’ willingness to pay and their ratings for the product quality. If the correlation is 1, meaning the willingness to pay and the ratings for the product quality are completely positively correlated and if the correlation is 0, meaning there is no correlation between these two variables. If the correlation is -1, it shows they are completely negatively correlated, meaning the higher one variable, the lower the other variable. If the absolute value of the variables is bigger than 0.5, they are usually significant.” – Jiafeng Li, Quantitative Data Analysis Techniques for Data-Driven Marketing, iAcquire; Twitter: @iacquire

  2. Predictive analytics shouldn’t discount human intuition. “The mobile marketer’s task, boiled down to basics, is to predict what the user or customer will do next, in order to position themselves in such a way as to be there, waiting for those anticipated user behaviors to happen, with a marketing strategy in hand.

“To accomplish this, it makes sense for marketers to hone in on characteristics and user behaviors that correlate, or seem to predict certain outcomes. If you’re collecting the data that make the most sense for your task, this is the first step in predictive analytics. Many marketing automation systems today are enabled with some predictive capabilities. Perhaps the most basic example is running and A/B test, showing two smaller segments different versions of a campaign, and then sending the best-performing variant to the rest of the user base. You’ve predicted what will perform better based on the data you have.

“But this doesn’t mean qualitative analysis is out the window. When we dig into our user data to look for patterns that will enable our predictive analytics tools, we must first ‘develop a qualitative understanding of conditions,’ says Nadeau. In less mathematical terms, this means using regular old human judgement to look at a scenario and parse ideas using only the power of the humble human brain.” – Lauren Leonardi, Data vs. Intuition: How to Approach Data for Great Marketing, Relate; Twitter: @Appboy

  1. Discriminant analysis is useful for a variety of needs. “This statistical technique is used to for classification of people, products, or other tangibles into two or more categories. Market research can make use of discriminant analyses in a number of ways. One simple example is to distinguish what advertising channels are most effective for different types of products.” – Gigi DeVault, Market Research 101: Data Analysis, The Balance; Twitter: @thebalance

  2. Know the difference between intelligent filtering and correlation. “Do learn the difference between intelligent filtering and correlation, as opposed to coincidence and co-visualization. Just because you have a picture of revenue performance against a depiction of buzz and sentiment, it does not mean the two have been intelligently correlated.” – Joshua Reynolds, Enough analysis, already! 8 tips for avoiding data fatigue, Martech Today; Twitter: @martech_today

  3. Understand standard deviation and why it matters. “The standard deviation, often represented with the Greek letter sigma, is the measure of a spread of data around the mean. A high standard deviation signifies that data is spread more widely from the mean, where a low standard deviation signals that more data align with the mean. In a portfolio of data analysis methods, the standard deviation is useful for quickly determining dispersion of data points.

“Pitfall: Just like the mean, the standard deviation is deceptive if taken alone. For example, if the data have a very strange pattern such as a non-normal curve or a large amount of outliers, then the standard deviation won’t give you all the information you need.” – John Dillard, 5 Most Important Methods For Statistical Data Analysis, Big Sky Associates; Twitter: @bigskyassoc

  1. Factor analysis is useful for analyzing large numbers of independent and dependent variables. “To analyze large numbers of dependent and independent variables, we might use factor analysis. This type of analysis can help detect what aspects of the independent variables are related to the dependent variables. When we receive the data, sets that are fairly wide, meaning that they had more variables in observations or records. We need a way to identify the core set of variables or drivers that will help to gain meaningful insight. Factor analysis can help identify that reduced subset of variables, meaning some of those variables represent similar relationships as those not included, but perhaps in a stronger way.” – Alex Mannella, PwC, Data-driven Decision Making, via Coursera; Twitter: @PwC

  2. Don’t discount the value of qualitative data. “There are differences between qualitative data analysis and quantitative data analysis. In qualitative researches using interviews, focus groups, experiments etc. data analysis is going to involve identifying common patterns within the responses and critically analyzing them in order to achieve research aims and objectives.

“Data analysis for quantitative studies, on the other hand, involves critical analysis and interpretation of figures and numbers, and attempts to find rationale behind the emergence of main findings. “Comparisons of primary research findings to the findings of the literature review are critically important for both types of studies – qualitative and quantitative.

“Data analysis methods in the absence of primary data collection can involve discussing common patterns, as well as, controversies within secondary data directly related to the research area.” – John Dudovskiy, Data Analysis, Research Methodology; Twitter: @Research_Net

  1. Inferential data analysis goes beyond descriptive analysis. “While descriptive data analysis can present a picture of the results, to really be useful the results of research should allow the researcher to accomplish other goals such as:
  • Using information obtained from a small group (i.e., sample of customers) to make judgments about a larger group (i.e., all customers)
  • Comparing groups to see if there is a difference in how they respond to an issue
  • Forecasting what may happen based on collected information

“To move beyond simply describing results requires the use of inferential data analysis where advanced statistical techniques are used to make judgments (i.e., inferences) about some issue (e.g., is one type of customer different from another type of customer). Using inferential data analysis requires a well-structured research plan that follows the scientific method. Also, most (but not all) inferential data analysis techniques require the use of quantitative data collection.

“As an example of the use of inferential data analysis, a marketer may wish to know if North American, European and Asian customers differ in how they rate certain issues. The marketer uses a survey that includes a number of questions asking customers from all three regions to rate issues on a scale of 1 to 5. If a survey is constructed properly the marketer can compare each group using statistical software that tests whether differences exists. This analysis offers much more insight than simply showing how many customers from each region responded to each question.” Step 6: Analyze Data, KnowThis.com; Twitter: @KnowThis

  1. Use cluster analysis to identify potential groupings that may not be obvious. “Cluster analysis is an exploratory tool designed to reveal natural groupings within a large group of observations, segmenting the survey sample – respondents or companies – into a small number of groups.

“Respondents whose answers are very similar should fall into the same clusters while respondents with very different answers should be in a different cluster. Ideally, the cases in each group should have a very similar profile towards specific characteristics – for example attitudinal or behavioral questions – while the profiles of respondents belonging to different clusters should be very dissimilar.

“Cluster analysis can suggest, based on complex input, groupings that would not otherwise be apparent, such as the needs of specific groupings or segments in the market.” Cluster Analysis, B2B International; Twitter: @B2B_Insight

  1. Logistic regression analysis can be used to make predictions of events. “Sometimes referred to as ‘choice models,’ this technique is a variation of multiple regression that allows for the prediction of an event. It is allowable to utilize nonmetric (typically binary) dependent variables, as the objective is to arrive at a probabilistic assessment of a binary choice. The independent variables can be either discrete or continuous. A contingency table is produced, which shows the classification of observations as to whether the observed and predicted events match. The sum of events that were predicted to occur which actually did occur and the events that were predicted not to occur which actually did not occur, divided by the total number of events, is a measure of the effectiveness of the model. This tool helps predict the choices consumers might make when presented with alternatives.” – Michael Richarme, Eleven Multivariate Analysis Techniques: Key Tools In Your Marketing Research Survival Kit, Decision Analyst; Twitter: @DecisionAnalyst

  2. Use data-over-time intervals. “Back in the summer, I activated the dormant Crazy Egg Facebook Page. Nothing too crazy, just posting our articles for our fans and occasionally a status update when I found something interesting to share.

“As the end of the year approaches, I would like to see if it makes sense to ramp up our presence on Facebook in 2013.

“But looking at data over time can be tricky if you don’t select the right intervals.

“In this case, going back 6 months and displaying information by the day is basically worthless. This view might be OK if I was trying to pinpoint the exact Facebook status updates that did well at driving traffic.

“But I just want to know if the posting to Facebook over the last six months has had any affect in the aggregate.

“When I view the data by the week, I get a much clearer picture of this data. This view would be perfect if I was, for example, trying to determine if a Facebook contest I ran generated more traffic than usual. But, again, that’s not what I am looking for today.

“Viewing the data monthly shows me that our small amount of activity on Facebook has produced more traffic. This view is (as Goldilocks would say) just right to answer this particular question.” – Russ Henneberry, See Google Analytics Data Clearly With These 3 Little Known Tricks, The Daily Egg; Twitter: @CrazyEgg

  1. Select the appropriate data analysis technique for the business case. “There are many well-developed methods available for conceptually or statistically analyzing the different kinds of data that can be gathered. When analyzing qualitative data, one can develop taxonomies or rubrics to group student comments collected by questionnaires and/or made in classroom discussions. The frequency of certain types of comments can be described, compared between categories, and investigated for change across time or differences between classes. Frequency data and chi-square analysis can supplement the narrative interpretation of such comments. For the analysis of quantitative data, a variety of statistical tests are available, ranging from the simple (t-tests) to the more complex (such as the use of factor analysis to develop scales).” Select Appropriate Data Analysis Techniques, MIT Teaching & Learning Lab; Twitter: @mit_tll

How to Get More Value from Your Marketing Data Analytics

  1. Assess your analytics capabilities, then fill in the gaps. “Marketing organizations have access to a lot of different analytic capabilities in support of various marketing goals, but if you’re like most, you probably don’t have all your bases covered. Assessing your current analytic capabilities is a good next step. After all, it’s important to know where you stand along the analytic spectrum, so you can identify where the gaps are and start developing a strategy for filling them in.

“For example, a marketing organization may already be collecting data from online and POS transactions, but what about all the unstructured information from social media sources or call-center logs? Such sources are a gold mine of information, and the technology for converting unstructured data into actual insights that marketers can use exists today. As such, a marketing organization may choose to plan and budget for adding analytic capabilities that can fill that particular gap. Of course, if you’re not quite sure where to start, well, that’s easy. Start where your needs are greatest, and fill in the gaps over time as new needs arise.” Marketing Analytics – What it is and why it matters, SAS; Twitter: @SASsoftware

  1. Use segmentation. “Segmentation not only allows you to separate the proverbial casual shoppers from the qualified leads, it also gives you details about your customers so that you can separate them into buckets and target them accordingly. By knowing who your customers are, where they are coming from, and what their purchase intent is, you’re better able to group them together and identify potentially unmet needs for distinct subsets of customers. These groups can consider anything from demographic info, to past shopping behaviors, to web usage patterns in order to make meaningful subsets out of your customers.” – Suzie Blaszkiewicz, Data analysis in marketing: a beginner’s guide, GetApp; Twitter: @GetApp

  2. Identify primary and secondary KPIs. “Don’t try to analyze it all; you’ll get lost in data and become discouraged and confused. Instead, narrow your focus to the metrics that will provide the most relevant insights.

“Having primary and secondary KPIs for your site will help you begin to narrow your focus. Common KPIs include conversion rate, revenue, CPA (Cost per acquisition), number of leads, and clickthrough rate.

“You will also want to study a few other secondary metrics that could affect your KPI.” – Lauren Pitchford, Marketing Analytics: 6 Simple Steps For Interpreting Your Data, Marketing Experiments; Twitter: @MktgExperiments

  1. Leverage predictive and prescriptive analytics. “Predictive & Prescriptive Analysis – in short, it is based on analyzing current and historical datasets to predict future possibilities, including alternative scenarios and risk assessment.

“Methods like artificial neural networks (ANN) and autoregressive integrated moving average (ARIMA), time series, seasonal naïve approach and data mining find wide application in data analytics nowadays.

“We’ve already explained them and recognized as one of the biggest business analytics trends for 2017. Your choice of method should depend on the type of data you’ve collected, your team’s skills and your resources.” – Mona Lebied, Your Data Won’t Speak Unless You Ask It The Right Data Analysis Questions, Datapine; Twitter: @datapine

  1. Recognize that more data won’t translate directly to positive, personalized experiences. “As consumers sign up, change, cancel email addresses, social media accounts, phone services and more, the dizzying array of consumer identifiers are constantly changing. This makes it much more difficult to accurately identify consumers to cultivate authentic personalized connections using dynamic and relevant content.

“Strength in numbers and scalability is a critical component of any brand’s marketing campaigns. However, inaccuracies in identity data, incomplete or incorrect lifestyle-indicating attribute data, or poor interlinking between them will significantly limit the reach of a campaign and weaken the effectiveness of personalization.

“As an example, if nearly 25 percent of targeted consumers use a different email for Facebook than what’s in a brand’s CRM system, a brand’s Facebook campaign will yield lower match rates, lower reach and less-than-optimal performance.

“Meaningful personalization requires real-time access to comprehensive, accurate, interlinked identity data at scale to ensure data is rich, accurate and complete, while also being able to suppress duplicate records and exclude those with a low propensity to buy. This identity data must be correctly interlinked with up-to-date attribute data such as consumer demographics and psychographics, mobile and online behaviors, propensity-to-buy scores and purchase histories through customized messaging that drives stronger response rates.” – Dave Dague, The game changer in retail personalization: Consumer identity management, Marketing Land; Twitter: @Marketingland

  1. Surface data doesn’t paint the picture you’re probably looking for. “Don’t be satisfied with ‘analyses of the average.’ If all you do is look at pie charts and bar charts of averages, are you really going to learn anything?

“One of my most profound business lessons came from a customer satisfaction report my company did every quarter. Instead of simply scanning through the bar charts and average customer satisfaction levels, I dug deep and read every comment, slicing the data different ways in an effort to find gaps and opportunities. There was one stray comment on one survey that was an early warning sign of a severe quality issue. It was so serious in fact, that one comment led to an investigation and a multi-million-dollar capital project.

“The real insight and innovation comes from the small data, not the Big Data.” – Mark Schaefer, Five lessons in marketing data analysis for beginners, MarkSchaefer.com; Twitter: @markwschaefer

  1. Integrate offline data by digging deep into past campaigns. “The first task is to look back at previous campaign reports. Not at what you thought was good, but what actually worked.

“Make a list of all the headlines you used. Then rank them by the simplest metric: whether a recipient responded. (Whether that means opening the mailer, or recalling it when you follow up by phone.)

“Then apply two more metrics: whether that recipient visited your website (you can put a unique URL in your letter to find out) and whether they went further down the sales funnel, such as providing their profile data at your squeeze page or landing page.

“Three valuable sets of data. And all came from an offline campaign.” – Julie Knight, Data Can Tell You What Makes Your Customers Tick, Marketscan; Twitter: @Marketscan

  1. Know the strengths and weaknesses of your data. “All data has strengths and weaknesses, and becoming familiar with them is essential. Spend extra time talking to experts — both technical and business — to not only reveal assets and liabilities, but also determine what the data does and doesn’t measure or capture. Data from marketing technology, for instance, measures what it was built for and not necessarily the entire market. Gain a sense in the beginning by working with the data to resolve simple questions — such as share of traffic by device — and comparing those answers to other sources to see whether they make sense. Not everything you carry out in this step will make the final analysis, but it will build invaluable knowledge and experience and prevent skewed numbers from entering the process.” – Anthony Power, Advanced Data Analysis — the Process Behind the Insight, Adobe Digital Marketing Blog; Twitter: @Adobe

  2. If analyzing a smaller, representative sample of data, employ sound sampling techniques. “When sampling, you need to decide what units (e.g., people, organizations, data) to include in your sample and which ones to exclude. Sampling techniques act as a guide to help you select these units. However, how units are selected varies considerably between probability sampling techniques and non-probability sampling techniques [see the articles, Probability sampling and Non-probability sampling to learn more about these types of sampling technique]. Moreover, there is also a lot of variation amongst non-probability sampling techniques in particular.

“Probability sampling techniques require a list of the population from which you select units for your sample. This raises potential data protection and confidentiality issues because units in the list (i.e., when people are your units) will not necessarily have given you permission to access the list with their details. Therefore, you need to check that you have the right to access the list in the first place.

“When using non-probability sampling, you need to ask yourself whether you are including or excluding units for theoretical or practical reasons. In the case of purposive sampling, the choice of which units to include and exclude is theoretically-driven. In such cases, there are few ethical concerns. However, where units are included or excluded for practical reasons, such as ease of access or personal preferences (e.g., convenience sampling), there is a danger that units will be excluded unnecessarily. For example, it is not uncommon when select units using convenience sampling that researchers? natural preferences (and even prejudices) will influence the selection process. For example, maybe the researcher would avoid approaching certain groups (e.g., socially marginalized individuals, people who speak little English, disabled people). Where this happens, it raises ethical issues because the picture being built through the research can be excessively narrow, and arguably, unethically narrow. This highlights the importance of using theory to determine the creation of samples when using non-probability sampling techniques rather than practical reasons, whenever possible.” Sampling strategies and research ethics, Laerd Dissertation

  1. Leverage data wrangling to generate the most robust data for analysis.Data wrangling, a core data analysis technique is not done in one fell swoop–it’s an iterative process that helps you get to the cleanest, most usable data possible prior to your analysis. Each step in the data wrangling process exposes new potential ways that the data might be “re-wrangled,” all driving towards the ultimate goal of generating the most robust data for final analysis.

“At Trifacta, we think about data wrangling process as the most critical first step and complimentary to other data analysis techniques.  Our data wrangling process includes six core activities to prepare data for analysis and to get the most business value out of your data:

  • Discovering – allows you to understand your data and how it’s useful for analytic exploration and analysis
  • Structuring – gives you the ability to format data of all shapes and sizes to work with traditional applications
  • Cleaning – lets you fix and standardize the data that might distort your analysis
  • Enriching – allows you to take advantage of the wrangling you’ve already done
  • Validating – identifies and surfaces data quality and consistency issues
  • Publishing – provides you the ability to plan for and deliver data for downstream analysis” Data Analysis Techniques, Trifacta; Twitter: @Trifacta
  1. Know how to leverage data and analytics to hack issues at the top, middle, and bottom of the funnel. “There are several ways to the hack middle funnel issues so let’s discuss a key method that consistently works well. First, measure your conversion points from the top of the funnel (i.e., web traffic) to middle of the funnel (i.e., lead/engagement). Second, measure the conversion from middle of the funnel to the bottom of the funnel. Measure your entire funnel conversion for at least two earlier time frames, preferably 3 months and 6 months in the past. Your historical funnel conversion metrics will become your benchmarks.

“Then, identify the gaps in current conversion and historical conversion benchmarks. Notice if there is a drop in conversion from traffic to lead gen or led gen to opportunity creation or product addition to the shopping cart. Once you have a laser target focus on the conversion impacting source you can drill down and separate technical and nontechnical issues. For example, a user can abandon the shopping cart as they are not able to update the products in the cart (technical issue) or they are not sure on the product prices (non-technical issue).” – Sameer Khan, Hack Low Performing Marketing Programs with Data and Analytics, KeyWebMetrics; Twitter: @SameerKhan

  1. Use a strategic approach to personalization. “Personalization is all the rage right now. And with good reason! This customer-centric strategy has major potential to motivate consumers to purchase. Plus, more and more experimentation platforms are enabling personalization.

“But the rise of personalization tools and popularity has meant the rise of marketers doing personalization the wrong way

“Before you start investing in visitor segments and personalized experiences based on guesswork and crossed fingers, consider the four most common personalization mistakes:

  • Ad hoc implementation of off-the-shelf tool features without understanding what need these features are solving.
  • Poor personalization insights with little data analysis and a framework that drives the implementation of the program.
  • Lack of a rigorous process to hypothesize, test, and validate personalization ideas.
  • Lack of resources to create the many additional marketing messages needed to support multiple, personalized target segments.” Chris Goward, Founder & CEO, WiderFunnel, as quoted by Sukh Dhillon, 4 Growth Hacking Tips from the Experts, Optimizely; Twitter: @Optimizely
  1. Once you have a platform in place for aggregating and analyzing data, you can begin to test. “Once you have a platform in place for aggregating and analyzing data, you can begin segmenting your audience and testing offers — essentially creating the foundation of a data-driven marketing strategy. Start by identifying personas, a characterization of ideal customer types, and then create a set of lead-generation campaigns to target them with compelling offers. Experiment with as many different marketing channels as you can to determine where you see the highest conversion rates.

“Once you take some baseline readings, you can begin the real work — testing and refining lead-generation campaigns over time to boost engagement and conversion rates. This will include making small changes as needed: writing a new subject line, changing the tweet’s hashtag, adding a video link, or changing your transmission time, etc. From there you can analyze lead results and use the intelligence to continuously optimize conversions from each and every campaign you run. You’ll be able to see what your audience is responding to as you continue to segment your target markets.

“For many, this is a complete mindset shift. But in the end, a data-driven marketing strategy is the fastest path to optimizing lead generation results, developing a loyal customer base and increasing revenue — the definition of growth hacking.” The Data-Driven Approach to Growth Hacking, Datorama; Twitter: @Datorama

  1. Use marketing data analytics as a decision-support tool. “Analytics are particularly valuable when used as a decision-support tool – a way of answering your key questions about what’s working, what isn’t, and what actions you should take as a result. To do this, he recommends following a simple, five-step process:
  • Formulate an idea about your content performance.
  • Determine a question you can ask to support this idea.
  • Create the report that will provide the appropriate data to answer that question.
  • Take action based on your analysis of that data.
  • Measure the results of the actions you take against the baseline data you gathered.” – Jodi Harris, Simple Tips for Sleuthing Your Site Performance Using Google Analytics, Content Marketing Institute; Twitter: @CMIContent

Presenting Your Findings in Meaningful Ways

  1. Use pie charts with caution. “The audience can only accurately gauge the size of pie slices if they are in familiar percentages (25%, 50%, 75%, 100%). Thus, it is difficult to compare other sizes effectively. Similar to pie charts, graphs with other special effects like 3D distorts the viewers’ ability to analyze size and length properly.” – Lynette Chen, 9 Tips for Building Actionable Data Visualizations, MassMedia; Twitter: @MassMetrics

  2. Add context to the data you pull from social media and other channels. “Social media be confusing sometimes, especially when you’re down in the day-to-day of it and struggling to see the bigger picture.

“Looking at the larger trends of your social performance is hugely helpful for seeing how strategies are playing out.” – Alfred Lua, 7 Social Media Analytics and Reporting Tips for Becoming a Data-Savvy Marketer, Buffer; Twitter: @buffer

  1. Show the likely cause with the effect. “Any time there’s a large spike or dip, it’s usually related to a change elsewhere in the system. Either a new referrer has appeared, an old one has disappeared, traffic from a different page has increased, or there’s been a decline in conversions from the previous step. A simple improvement would be to show the likely cause with the effect.” – Des Traynor, Four Things I Wish Every Chart Did, Intercom; Twitter: @intercom

  2. Keep data presentations (whether visualizations, slides, written reports, etc.) as simple as possible. “Bring insane focus to your data presentation. If you can, focus on a singular metric for each module/slide/element. Then present the data as simply as you possibly can. And often, you don’t need to go very far from the defaults in Excel – though you are welcome to use any software you want.” – Avinash Kaushik, 7 Data Presentation Tips: Think, Focus, Simplify, Calibrate, Visualize++, Occam’s Razor; Twitter: @avinash

  3. No matter what the results, present data analysis findings truthfully. “Believe it or not, it is possible to be in marketing and to use data in both meaningful and truthful ways (i.e. the good). Unfortunately, this means that you need to be prepared to accept and present results that don’t necessarily tell the story you want it to. Whether it’s poor or underwhelming performance or an outcome that doesn’t validate your hypothesis, having the bravery to present these results to your leadership, your team or your client can be the difference between facing a hard truth but making a good decision and wasting time and effort on producing fancy tables and charts.” – Stephen Tracy, Meaningful Interpretation Of Data: The Good, The Bad And The Ugly, Analythioal; Twitter: @stephen_tracy

  4. Guide the conversation. “’Let the data speak.’ It’s a common saying for chart design. The premise — strip out the bits that don’t help patterns in your data emerge — is fine, but people often misinterpret the mantra to mean that they should make a stripped-down chart and let the data take it from there.

“You have to guide the conversation though. You must help the data focus and get to the point. Otherwise, it just ends up rambling about what it had for breakfast this morning and how the coffee wasn’t hot enough.” – Nathan Yau, One Dataset, Visualized 25 Ways, FlowingData; Twitter: @flowingdata

  1. Choose the right type of chart or graph to present the data. “If you want to know more information about how a data set performed during a specific time period, there are specific chart types that do extremely well.

“You should choose a:

  • Line
  • Dual-Axis Line
  • Column

“Relationship charts are suited to showing how one variable relates to one or numerous different variables. You could use this to show how something positively effects, has no effect, or negatively effects another variable.

“When trying to establish the relationship between things, use these charts:

  1. Set your expectations. “You will never be able to visualize every single action that has ever been taken on the site in a single chart. You can, and should, be filtering, segmenting, grouping, aggregating, and doing everything you can think of to get a simplified view. Yes, this absolutely means that many of the details will be hidden.

“Are you trying to answer a specific question or are you hoping for a golden nugget of insight? Hint – the more specific your questions get, the easier they are to answer. Know that going into this exercise.” – Becky West, Visualizing the Customer Journey with Google Analytics Data, LunaMetrics; Twitter: @LunaMetrics

  1. Focus on the story. “Data analysis isn’t about graphics and visualizations; it’s about telling a story. Look at data the way a detective examines a crime scene. Try to understand what happened and what evidence needs to be collected. The visualization—it can be a chart, map or single number—will come naturally once the mystery is solved. The focus is the story.” – Daniel Waisberg, Tell a Meaningful Story With Data, Think with Google; Twitter: @danielwaisberg
  2. Visualize and present data with the audience in mind. “Plenty of people can create a bar chart, but what sets apart good visualizations from the great ones are those created with your audience in mind—not a one-size fits all. What is important to executive leadership may be different from the marketing team. Visualizations should also differ based on your audience’s familiarity with data. A group that is well versed on the local market and its trends can handle more complex visualizations that dive into the details. This same group may be more interested in quarterly trends. While the group that is less familiar may require more simple, high-level information to get them acquainted with the market.

“Marketers need more detail on the ‘who’ aspect of the data. They need the data translated to a format that turns numbers into people. While it is important for them to know high-level market trends, such as a loss in cardiology market share, it is more important for them to understand who those lost cardiology patients are—where do they live, how are they insured, how do they spend their free time, what is their age range, where do they shop, and who are they utilizing for care.” – Morgan Atkins, Pro Tips for Presenting Market Data Visualization to Any Audience, Stratasan; Twitter: @stratasan

What’s your go-to tactic for marketing data analysis that you can’t live without?

  •  
  •  
  •  
  •  
  •  
Posted in Big Data, Marketing

Leave a Reply

Your email address will not be published.