in , ,

How to Use Data Science to Improve SEO

how-to-use-data-science-to-improve-seo

Optimizing a website for search traffic can take time. As Google’s algorithm can be unpredictable, it’s not a simple or straightforward task and SEO involves having to make estimated best guesses.

As SEO professionals, the more information you possess, the better you can make predictions and see results from your SEO strategies. By using data science, you can get valuable insights into your website’s performance and answers to some of your SEO questions. Data driven information will help you make the right decisions and improve your optimization campaigns.

In this SEMrush Chat, Qi Zhao, our data scientist discussed with the community how data science can impact SEO. This post is a summary of tips from the industry on how using the right data can help you achieve SEO success.

WebCEOSEO Tools

First of all, you need to decide where you will get your data from. We asked our guests to name marketing analytics tools they use to make their SEO more predictable and quantifiable.

  • Excel. Even though not everyone likes Excel, it’s still one of the easiest and fastest ways to manipulate a large amount of information.
  • BigQuery. This big data analytics platform is used for processing large datasets. The service runs on the Google Cloud Storage infrastructure.
  • Tableau. This company produces interactive data visualization software, which allows you to transform data into appealing, interactive visualizations.

A1: In SEMrush, we use Excel, Google Analytics, R, Bigquery, Tableau (visualization).And @semrush! Any other recommendations? #SEMrushchat

— Qi Zhao (@QiZhaoDATA) July 13, 2016

  • Google Analytics. This is probably one of the most popular tools for measuring SEO success. Dallas McLaughlin described five effective Google Analytics strategies for measuring search engine optimization success in his article posted on Kissmetrics.
  • Google Search Console. Search Console is a free service that helps users monitor and maintain their website’s presence in Google search results. It explains the fundamentals of search and helps you understand how Google sees your site.
  • These two services provide necessary data on what is working and what is not in your SEO campaign, as Reva Minkoff ‏@revaminkoff points out.

A1: Google Analytics, Search Console, Bing Webmaster Tools, Screaming Frog & Moz Analytic some more SEO than strictly Analytics #semrushchat

— Annalisa M. Hilliard (@ahilliardm) July 13, 2016

  • Google Keyword Planner. This new service, which replaced Google’s previous Keyword Tool, has a whole set of features that help users identify a keyword’s increasing or decreasing popularity, filter out low-volume searches, and easily create multiple keyword combinations.
  • Keywordtool.io. Keyword Tool is a good alternative to Google Keyword Planner. The free version of this tool will generate up to 750 long-tail keyword suggestions for each search term.

a1) SEMRush, KeywordPlanner, Google Analytics – start by understanding audience behavior and you won’t be surprised #semrushchat

— Ryan Johnson (@rsj8000) July 13, 2016

  • Screaming Frog. Screaming Frog is used by SEO agencies and independent SEO consultants around the world. Its SEO Spider Tool will crawl a website that you want to check and examine its URLs for common SEO issues.

A1 Screaming Frog provides some helpful insights about how a site is set up that are worth peeking behind the curtains for. #semrushchat

— Bill Slawski (@bill_slawski) July 13, 2016

  • Social media analytics. Rachel Howe made a great point by adding social media analytics to this list. With the significant rise of social media, companies all over the world are beginning to take advantage of the opportunities that social networks provide. Analyzing data gathered from various social media websites can help businesses make the right decisions.

A1: Social analytics is important too if you’re trying to get better results from distribution – Twitter, FB, Pinterest #semrushchat

— Rachel Howe (@R8chel_Marie) July 13, 2016

Data Science and SEO Q1. recap:

Using these tools, you can leverage SEO best practices and improve your website’s performance in organic search:


Google Keyword Planner, Google Analytics, Adwords, SEMrush, ScreamingFrog, Excel, Majestic, Powersuggest, RavenTools, Searchmetrics, Google Search Console, BigQuery, Tableau and Keywordstool.io

Get a free trial of Traffic Analytics

Get a full view of any company’s online performance

Please specify a valid domain, e.g.,
www.example.com


One of the key goals of any marketing campaign is to align SEO best practices with other marketing channels in order to better leverage all marketing activities. Let’s figure out how to use data science to accomplish this task.

Qi recommends identifying your dependent and independent variables, also known as outcome and predictor variables. Dependent variables hinge on other factors – independent variables. Asking the right questions will lead you to the right answer.

In the context of business, a variable is usually an event, a value proposition or some other thing that a company can measure. Understanding how to use the interplay between these two types of variables to measure the effectiveness of different marketing strategies can help your company maximize its marketing impact.

A2: Choose your outcome, predictor variables, do the maths, and visualise your findings. Create matrix for all campaigns #semrushchat

— Qi Zhao (@QiZhaoDATA) July 13, 2016

Sharpe Digital suggested using data to access your website’s usability. You should check your site’s navigation and improve it, if necessary. By creating a smooth user experience, you’ll ensure that your visitors can easily browse through your site’s pages and that they’ll want to stay on it.

Knowing your goals and your audience will influence your decisions and help you create more effective marketing strategies in the future. Your marketing campaigns should be based on your audience’s needs, according to Obed M. ‏@MrClics: “Make better decisions and align your campaigns with the needs of your audience.”

ThinkSEM added that, today, it’s really difficult to achieve success online without knowing how to use data analytics: “Frankly, if you don’t know how to use data/analytics, you’re dead in the water.”

By using trend graphs, you can identify some of your website’s pain points.

A2. Use trend graphs to help identify pain points, exit traps and buyer preferences. Identify outliers and their motivations #semrushchat

— Chris Desadoy (@EliteYouTubePro) July 13, 2016

Clients interact with your brand in multiple ways. Bill Slawski mentioned several important metrics you should keep track of: social interactions, mentions, reviews and sentiments. Use these metrics to understand the relationship between your customers and your company.

A2 Track and monitor things such as social interactions/mentions/reviews/sentiment in ways that you can take action on. #semrushchat

— Bill Slawski (@bill_slawski) July 13, 2016

  • Use trend graphs to help identify pain points, exit traps and buyer preferences. 
  • Reveal the best sources of traffic and compare which kind of traffic leads to more engagement.
  • Track everything.
  • Remove friction to increase CTRs/conversions.
  • Identify your ‘actual’ vs. your ‘targeted’ audience and adjust to make them become one and the same.
  • Show how other channels work alongside SEO – attribution and multi-channel funnel reports.
  • Demonstrate the ROI of SEO – goal conversions based on rankings, paid search $ equivalent of organic traffic.
  • Look for low-conversion pages that rank well and get a decent amount of organic traffic and optimize for some quick wins.

Interpreting your data correctly is key. By using data science appropriately, you can make improvements to your website in order to satisfy your audience’s needs and wants.


Server log analysis is an excellent way to improve your technical and on-site SEO. Going through your log files can help you better understand search engine crawlers behavior for your website. A server log file contains a record of each visit to your website visit and can reveal information that is crucial for thorough SEO analysis.

Many people turn to log analysis only in response to problems, like bugs or errors, or because they’ve been hacked. Nevertheless, it should be done on a regular basis.

A4: In my experience, fail to frame data in a way that addresses the question seeking to answer. SYNC the data to the question.#SEMrushchat

— Qi Zhao (@QiZhaoDATA) July 13, 2016

Our chat participants mentioned several issues that SEO specialists should definitely pay attention to when they perform log analysis. It helps you identify your website’s architecture (internal link structure); duplicate, low-quality, or thin content; and your site’s performance.

A3: Server log analysis can help to identify internal link structure, duplicate/ low quality content & site performance/ speed #semrushchat

— WooRank (@woorank) July 13, 2016

Analyzing your log files can point out URLs that are not often crawled by search engine bots. You may even find that Google is not crawling or indexing one of your website’s crucial pages. Your logs will show you which URLs and directories are getting the most and least attention from the search engine.

A3: Most vs least crawled URLs – see where search engines are spending the majority of their time when crawling your site. #semrushchat

— Omi Sido (@OmiSido) July 13, 2016

It’s also possible that bots are following or indexing pages that you don’t want them to follow or index. Also, your website will lose potential traffic if some of its important pages are not indexed.

A3: Crawlers following they’re not supposed to follow. Or indexing pages they’re not supposed to index. 404 Errors as well. #semrushchat

— Hernan Vazquez (@HernanV07) July 13, 2016

Hernan Vazquez also suggested checking your sites load time. Internet users are becoming more and more impatient. A short page loading time is a vital part of your website’s user experience. People will leave your website if it takes too long to load.

According to Kissmetrics, 47 percent of visitors expect a webpage to load in two seconds or less.

Andy Drinkwater points out that not all people set specific goals before performing server log analysis.

A3: Not deciding on their goals before they delving into the murky water that is a server log #semrushchat

— Andy Drinkwater (@iqseo) July 13, 2016

He also thinks that some SEO specialists don’t know how to look for leaks or missed opportunities where crawl budget is being wasted.

  • Check loading times, indexing, bounce rate, 404 pages, response errors and redirects.
  • Look for leaks or missed opportunities where crawl is being wasted.
  • Issues: crawlers following what they’re not supposed to follow, duplicate crawling or indexing pages they’re not supposed to index.
  • Look for unusual through traffic.
  • You need to be able to tell if you’re being used a redirect from your logs.

Server log analysis is an accurate way to find out how search engines crawl your website. Look out for the problems that our guests mentioned when performing log analysis in order to eliminate crawling issues that might be weakening your SEO efforts.


To reach your target market, it’s important to not only gather data, but also to gather the right data. Companies should pay attention to the accuracy of their analytics in order to make sure that their data insights can help them achieve their desired results.

Our chat participants shared their insights on how SEO specialists can improve the accuracy of their big data analysis.

Qi Zhao thinks that one of the biggest challenges for SEO specialists and digital marketers today is that they fail to frame their data in a way that addresses the question they are seeking to answer. Because they have to deal with vast amounts of data, they must learn how to analyze it in order to leverage their strategies. Our special guest recommended that SEO specialists and marketers synchronize their data with their questions.

A4: In my experience, fail to frame data in a way that addresses the question seeking to answer. SYNC the data to the question.#SEMrushchat

— Qi Zhao (@QiZhaoDATA) July 13, 2016

Many of our guests warned against focusing on just one source. “Never trust just one source. The more sources, the better the results,” says Andy Drinkwater ‏@iqseo. If you can’t understand your data and adjust your marketing strategy accordingly, it won’t be beneficial for you.

It’s important to find the right tools to help you interpret big data. “Use multiple tools to validate your data sources and get access to AdWords data,” as Arnout Hellemans ‏@hellemans recommends.

A4) Don’t trust just one source. Understand tools’ weaknesses (e.g. what they don’t track). And take a break to get perspective #semrushchat

— David Rosam (@writingforseo) July 13, 2016

In addition to using the right tools, you should pay attention to the right analytics. Focus on real analytics, as Omi Sido advised. Your metrics and customer actions should be directly tied to KPIs that reflect conversions and customer loyalty. Focus on actionable metrics, rather than vanity ones.

A4: Continually focus on real analytics – the ones related to your KPIs that reflect conversions and customer loyalty.#semrushchat

— Omi Sido (@OmiSido) July 13, 2016

It’s important to not only know your SEO campaign’s numbers, but to be able to tell the story behind those numbers. Data alone can’t always help you take action. Good data is most valuable when it can be comprehended by decision makers. David Kutcher recommends focusing less on numbers and more on stories and the behaviors behind them.

A4: assume a level of inaccuracy, then focus less on minute numbers, more on the stories and behaviors behind them. #semrushchat

— David Kutcher (@confluentforms) July 13, 2016

Any individual, no matter how efficient, professional and capable they may be, can make a mistake. Agent Palmer reminded us that a good way to overcome data inaccuracy is to have more than one person analyze your data.

A4: The best analysis to counteract inaccuracy is to have more than one person doing analysis and have them TALK to each other. #semrushchat

— Agent Palmer (@AgentPalmer) July 13, 2016

Our guests recommend paying attention to real analytics, focusing less on numbers and more on the stories behind those numbers, choosing the right tools and relying on several sources instead of just one.

  • It’s all about money and not rankings. Focus on what matters.
  • Use multiple tools to validate data sources.
  • Focus less on minute numbers and more on the stories and behaviors behind them.
  • Never trust just one source. The more sources, the better your results.
  • Understand tools’ weaknesses (e.g. what they don’t track) and take a break to gain perspective.
  • Continually focus on real analytics – the ones related to your KPIs that reflect conversions and customer loyalty.

What have you done to improve the accuracy of your big data analysis? We would love to hear about your experiences!


Data analytics provides companies with a wealth of information that can be considered when it comes time to make critical marketing decisions. But how can marketers determine which data are the most crucial. At the end of our chat, we asked our guests what data every marketer should analyze to ensure their marketing activities deliver better results and help them achieve their business goals.

First of all, it’s important to clearly understand your business goals. There is absolutely no point in analyzing the wrong metrics.

A5: Understanding what the business goals are. There is no point in chasing stats if they are the wrong ones #semrushchat

— Andy Drinkwater (@iqseo) July 13, 2016

Our special guest reminded us of the importance of customer segmentation. It allows you to focus on specific groups of people to target. Better segmentation leads to better targeting, higher conversions and higher profits.

Customer retention is important for any business. It’s important to determine how good your customer service is and how quickly you can grow your business.

Qi Zhao also mentioned churn prediction, the process of identifying customers who are likely to cancel their subscription to your service. By predicting customer churn early and responding to it quickly, you can keep more of your existing customers.

A5: Conversions matrix, Customer segmentation, Retention & churn prediction. #semrushchat

— Qi Zhao (@QiZhaoDATA) July 13, 2016

Some of our chat participants pointed out that the primary goal of any business is a higher profit. Your conversion rate is crucial, because it’s the metric by which you can measure your business’ profitability. You can use it to predict your future success or understand what’s working and what isn’t. Nevertheless, analyzing your conversion rate alone won’t show you the whole picture of your website’s performance.

A5: You should be taking into account business goals for your SEO goals. All businesses want 2 make more $. Focus on conv rate #semrushchat

— Rachel Howe (@R8chel_Marie) July 13, 2016

Dawn Anderson ‏@dawnieando shared her experience and said that she is now focusing more on conversion, retention and customer lifetime value (CLTV): “After years of chasing traffic, I’ve decided to focus a lot more on conversion, retention, CLTV.”

Understanding CLTV involves calculating how much net profit a customer will generate over the full span of their relationship with your company. CLTV provides valuable insights into your business and helps you optimize your marketing budget.

Chris Desadoy suggested keeping track of your first-time visitors conversions and the conversion rates of your returning or loyal clients. The way a new visitor will interact with your website will differ from way a returning visitor will interact with it.

A5. Should look at new vs returning, time on site, shares per page, incoming traffic, outbound traffic and custom events #semrushchat

— Chris Desadoy (@EliteYouTubePro) July 13, 2016

Finally, the web traffic sources metric measures which traffic sources are driving visitors to your site: direct, referral, organic, and paid. By analyzing each traffic source, you can find out more detailed information about your web traffic. If you understand how users are finding your website, you will be able to effectively optimize your marketing budget, maximize its impact, and draw more customers.

A5 – content that is driving interest, conversions, buyer/customer touch point analysis, VoC, traffic source analysis, etc. #semrushchat

— Sue Duris (@SusynEliseDuris) July 13, 2016

What data should every marketer analyze to ensure their process helps them achieve their business goals?

  • Conversions
  • Traffic sources
  • Traffic quality
  • Revenue
  • Customer lifetime value
  • Visitor path to see where drop-offs are occurring.
  • Traffic is useless unless your visitors become customers.

Data science provides more accurate answers than any non-data-related science. It changes the way we make decisions and help us discover the latest trends. It’s all about answers and decision-making. But having more data is only better if you know what to do with it.

Thanks to all our guests for their valuable tips and great insights into data science and SEO!

Get a free trial of Traffic Analytics

Get a full view of any company’s online performance

Please specify a valid domain, e.g.,
www.example.com

What do you think?

WebCEOSEO Tools

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *

Loading…

0
WebCEOSEO Tools
bing-site-explorer-has-unique-tools-for-seos-via-@mattgsouthern

Bing Site Explorer Has Unique Tools for SEOs via @MattGSouthern

how-to-improve-your-local-seo-in-5-steps

How to Improve Your Local SEO in 5 steps