How ecommerce companies can use data for better decision-making

How ecommerce companies can use data for better decision-making

Data is now the most valuable resource on the planet.

If you’ve read any of our other recent blog posts, you’re probably aware of the fact that data recently surpassed oil as the most valuable resource on Earth. While that came as a shock to some, to others this has been a long time coming.

Studies show that data-driven businesses are 23 times more likely to acquire customers, 6 times as likely to retain those customers, and 19 times as likely to be profitable.

As businesses have realized the value of data, the demand for more and more consumer data has exploded. Despite the general acknowledgement of the value of data, it’s estimated that 60-73% of data collected isn’t used in decision-making.

In this post I’ll cover a couple of ways that you can leverage data to make better decisions in your ecommerce business.

Understand your customers

Most marketers understand the importance of using data to drive their marketing decisions. The problem that most marketers face is getting accurate data that they can trust in order to make the right decisions. So that’s where we’ll start.

Overattribution

Truly the bane of every marketer’s existence, over-attribution is a constant in today’s marketing landscape. An example of over-attribution would be when you look at Facebook and they claim to have generated $10K in sales, and then you look at Google and they claim to have created $10K in sales, but you only had $15K worth of sales in that period.

Over-attribution occurs for a myriad of reasons. One of the primary reasons that it can occur is that the different ad platforms utilize different conversion reporting. Facebook currently utilizes a 28-day click and 1-day view conversion window. That means that if someone clicked on your Facebook ad and then came back and purchased from you within 28 days, they claim 100% responsibility for that sale. Google, on the other hand, utilizes a last-click attribution model. That means that they award 100% of the credit for the sale to the last click that someone used before purchasing.

UTMs

There are many solutions to solving over-attribution, but none are perfect. The first solution that we always recommend is UTMs.

UTMs are pieces of tracking information that you can append to a URL in order to improve your tracking. These can help you see exactly what ads, emails, or blog posts people clicked on in order to get to your site.

UTMs are amazing for increasing the granularity of your tracking and allow improved insights into what efforts actually drove people to your site. Unfortunately though, they don’t completely solve the issue of over-attribution. While they will allow you to see exactly what ads drove people to your site, you still have to deal with the different attribution windows in your reporting.

Multi-touch attribution

The best solution to the over-attribution problem is, unfortunately, also one of the more complicated ones. Multi-touch attribution most accurately reflects the client journey across platforms. By tracking the clients journey, these models can assign a portion of the total sale revenue to each platform that took part in the client’s journey. The reason that these can get complicated is because you need to model and decide how you want to assign credit to each platform.

Some of the more popular models that people use are: time decay, which allows you to decrease the amount of credit given to each touch point based off how long ago that happened; position based, which assigns 33% of the credit to both the first and last touch points, and then distributes the remaining 33% equally across the other touch points; the final option that we want to cover here is linear, which just assigns equal weight across every touch point.

Both UTMs and multi-touch attribution have their place in a marketers tool chest. We always recommend using UTMs, and multi-touch attribution can help with more advanced marketing initiatives.

Purchasing behaviors

Once you know where your customers come from, the next thing that you need to know is what they’re buying from you. Thankfully most ecommerce platforms readily provide this information. The important metrics to look at here are: average order value (AOV), lifetime value (LTV), and repurchase rates. Additionally, you should examine each of these metrics through the lens of how different products affect them.

In the early stages of a business, AOV is extremely important. We’ll cover more on this later, but the important thing to note is that if you can keep your cost per acquisition (CPA) below your AOV then you’ll always drive a profit off your ads. This will allow you to scale your advertising, and your company with it.

As you grow more advanced in your tracking and data, LTV becomes more and more important. As you grow in your understanding of LTV, AOV begins to matter less. Rather than worrying about driving a profit off the initial purchase, you can take a loss up front. Knowing the lifetime value of your clients gives you more freedom and flexibility in the acquisition of clients. This can lead to explosive results, just see what it did for Danette May:

The final important metric that you need to know about your customers ties in with AOV, and that’s repurchase rates. If you know when your clients will come back and repurchase from you, then you can accurately chart how long it will take for you to break even on your ads. Even more importantly, charting this metric over time allows you to see how your post-purchase marketing efforts affect your customers.

Understand your costs

In addition to understanding your customer behavior, you need to understand your operational behavior. We talked a lot about acquisition costs and advertising costs in the previous sections, but another important cost is the cost of goods sold (COGS).

In order to determine an acceptable CPA, you need to know what the costs of your business are.

Every business has their own view on how they calculate this metric. Some choose to include their operational costs in their COGS. Some only roll in the marketing costs, but not the salaries of the team. You need to determine the costs associated with the products that you sell in order to properly decide on acceptable margins.

Once you know the margins that your business needs in order to operate properly, then you can appropriately decide on your allowable CPA.

Tracking these metrics will allow you greater insight into your business and customers. Armed with this data, you can create exponential growth.

Praxis Metrics data ownership

The importance of data ownership

In this episode of the Data Rich show, AJ is joined by Kevin Brkal the president and founder of KNB Online Inc.

We talk through attribution models, ad spend, and how to protect your data through data ownership.

Check out our insights and conversation below:

Data- Amazing, but creepy

Kevin’s agency focuses on Facebook ads. The reason that they chose this as their platform is because of the Facebook audience network. Lots of different sites use the Facebook Ads network to sell ad space on their websites.

When it comes to the mobile web, any apps that collect real time data most likely use or sell that data. They can track your location and establish geo-fencing and geo-targeting to hyper target you as a consumer. It often freaks people out when they start to see ads for things that they think shouldn’t have a digital trail, but any number of apps on your phone could theoretically track that information on you.

As people get more and more creeped out by the things about them that are tracked, we see platforms cracking down on the things that can be tracked. The question that we naturally want to answer is how will this affect marketing.

Find a strategy that works for you

Before the internet, marketers still reached their target market. While we may not have access to as much information as before, you can still get mountains of data.

While browsers crack down on the data that marketers can access, no one is looking at location data being shared. Search engines also still sell randomized user data as well, so marketers shouldn’t panic just yet.

The primary victim of the data crackdown

Attribution modeling will get more difficult as browsers cut down on the amount of data that they share. This leaves marketers to rely more heavily on last-click attribution, or just saturate their markets with ads. The business who can afford to spend the most to acquire their customers always wins, but this may become even more important in the future.

Attribution is already a mess, but as browsers continue to limit the amount of data that you can gather on customers, it will only get worse. This change increases the confusion that ecommerce companies will have to deal with. Businesses just need to just gather as much data as possible to make an informed decision.

The solution:

As we talked about previously, the best way to combat the confusion is to gather as much data as possible.

One of the best way to increase your data is to leverage UTMs. UTMs are free tools that everyone can use to increase the amount of data that you can gather. They allow you to create custom tracking parameters to gain better insights into your customers.

From there, you need to track your data in as many sources as possible. Facebook Pixels, Google Analytics, the back end of your ecommerce store; all of them track data differently. But if you have all of that data tracked, then on the back end a data company can extract the data and figure out the truth for you.

Another thing that you can do is alter the attribution models that you use in your tracking systems. Facebook defaults to a 28-day click window, and a 1-day view-through window. You can alter this window to better match your preferred attribution modeling.

Attribution modeling

In today’s marketing landscape, there is no limit to the number of touch points that you can have with your customer. The trouble that most businesses run into is deciding how they want to attribute back to the touch points across the journey.

Google Analytics defaults to a last-click attribution model. This means that the thing that drove them to your site the time that they converted gets complete credit for the sale. Facebook has an attribution window, in the which it claims full credit if a sale occurs in that window. The trick that marketers need to use is a blended model.

Adidas recently stopped their branding campaigns in favor of campaigns that seemed to be driving their sales, based off last click attribution models. They quickly discovered that the branding campaigns that they were running warmed their customers enough to click on the direct response campaign and purchase. Based off the data that they looked at, they thought that they made the right choice to cut the branding campaigns.

Data ownership

The most important thing that companies need to do when working with an agency is make sure that you’re owning your data. Whatever agency you work with, you need to make sure that they create the ads under your account and that they track with your pixel. The reason for this is that then you own that data. If you allow them to run the ads under their umbrella or with their pixel, then they own the data. In the event of a dissolution of the partnership, they could sell that data to your competitors.

Regardless of who you work with, it’s extremely important to own your data. Many ecommerce companies have started to move away from Amazon, because Amazon owns all of the data on its platform. As Kevin correctly pointed out, Amazon can take the data that they gather on your customers, and the things that they purchase from you. From there, they can recreate your product under their umbrella and force you out of your own market. It wouldn’t be the first time that they did.

Businesses have finally begun to recognize the value of data, as data has just surpassed oil as the most valuable resource on the planet. Businesses have begun to recognize the value of owning their data from top to bottom.

Take action from your data

Once you own your data, the next important thing to do is take action from it.

In order to know what actions to take, you need to know what your primary objective is. If you’re an ecommerce company, you likely want to increase sales. B2B companies likely want to increase leads. The important metrics that everyone should track are the cost per acquisition (CPA) or cost per lead (CPL). From there you want to calculate your return on ad spend (ROAS). If you’re looking at generating leads, it’s important to know how much it costs for you to turn a lead into a customer, or your conversion rates from leads to customers. Once you know that, you can find the lifetime value of those customers (LTV). If you have your LTV and your conversion rates, then you can reverse engineer your allowable CPL.

Don’t take too quick of action though

With all data, it’s important to find as many sources of validation as you can. In this hyper-connected world, it’s unfortunately easy to have skewed data; as this commercial points out:

Many analytics systems fail to recognize refreshes on thank you pages, which can dramatically skew your data. At Praxis, we have found that the best way to stop this is by creating a first party cookie that loads into the user’s browser the first time that they visit the thank you page. From there, you can update your tags to only fire in the event that the cookie is not present in the browser. Obviously, this still isn’t a perfect solution to the problem, but it can reduce the negative effects of over-attribution.

The most important metrics

Kevin has found ROAS to be the most important KPI in his business. Anyone who runs ads obviously wants to turn a profit on those ads, so it is very important to make sure that you track your ROAS.

Regardless of the metrics that you measure, the most important thing that you can do with data is validate. Track everything through multiple systems, don’t put your trust in any one. At the end of the day, these platforms want you to spend more money with them, so they will skew the data in their favor.

Data can be overwhelming at first, but it is your friend. Data will help your business scale and grow faster than anything else. AJ pointed out that you need a relationship with your data. The more time you spend with your data, the more in sync you can get with it. The better attuned you are to shifts in the data, the faster you can react to sudden changes or opportunities.

If you need help with Facebook advertising, Kevin is a great resource and can be reached at kevinbrkal.com

Praxis Metrics- Leveraging data to optimize ad spend

Praxis Metrics – Leveraging data to optimize ad spend

In this guest appearance on the Perpetual Traffic podcast, AJ and Meaghan talk about how to use your data to optimize your ad spend, and rapidly scale your business.

They cover everything from getting your tracking in order, all the way up to creating customized dashboards and leveraging complex machine learning and AI.

Enjoy the episode and our insights below.

The struggle today-

Many marketers feel that they aren’t getting the most accurate data inside of the ad platforms. Unfortunately, they are completely correct. Some marketers go so far as to purchase a cheap dashboarding tool in order to help them bring all of their metrics together into one platform in order to help them with this issue. Unfortunately, this will not solve the problem for them at all.

Why do we suddenly have this struggle with data? What drove us to this point?

In our opinion, the problem stems from an overabundance of data. Never in the history of the world has so much data been available to us. Even in the last 20-30 years, large-scale data projects were reserved exclusively for enterprise-level companies. But now, every company has access to “big data”; despite this, many still have the mentality that their business doesn’t have the same access to data, and therefore the same opportunities and responsibilities, as the larger organizations.

Because these smaller businesses fail to leverage the data available to them, they often find themselves utilizing incomplete or dirty data. If they utilized all of the tools and tracking options available to them, they would have a much more complete and accurate picture of what’s happening.

The opportunity today-

Similar to the dot-com boom of the late 90’s, we’re seeing a “data boom” today. Those that have embraced data and created strategic initiatives around data are already separating themselves from their competition. Taking action from data is the new competitive advantage.

Those who capitalize on data have the opportunity to outpace and out-scale their competitors. John Wanamaker said: “Half the money I spend on advertising is wasted; the trouble is I don’t know which half.”. Those who eliminate waste in their budget open themselves to amazing opportunities. By doubling or quadrupling down on the things that work, they can drive exponential growth.

How you can capitalize on this opportunity-

Ask the right questions-

We firmly believe in the Socratic method. Asking questions helps you find deeper truths. The trick is finding the right questions to ask that will propel your business forward.

We found that the best way to find these questions is through a process called “metrics mapping”. The diagram below walks through an example:

Praxis Metrics- Metrics Mapping

Metrics Mapping starts with the big goals of your organization. This could include doubling your revenue year over year, increasing sales of a certain product by 30%, etc. From there, we want to drill down to the questions that you need to answer in order to meet that goal. If you want to double your revenue, then why don’t you? What questions do you need to answer in order to hit your goal?

Once you have the questions that you need to answer, it’s time to figure out what numbers can help you answer that question. In the example above, we need to know how to increase conversions and revenue from the website. In order to figure out how to do that, we need to figure out conversion rates, LTV, CPA, and profitability.

Once we have asked the right questions and gathered the necessary data, we need to:

Get granular with it-

Averages are inherently evil. Averages by definition mash together your highs and your lows and give you one number to work with. In order to properly scale your business, you need to know what creates the highs and what creates the lows. Once you know that, you can scrap the things that bring in the lows and double down on the highs.

Going back to the previous example; once we’ve gathered the numbers, we can strategize our next move. Perhaps we need to update our nurture sequence to increase return purchases. We may have a funnel step that causes dramatic drop-off that we can eliminate.

By getting granular in our analysis, we can discover a myriad of opportunities.

What metrics should every business look at?

Every business suffers from “terminal uniqueness”. While every business has certain things that they specifically need to track, there are a host of metrics that every business should know.

The obvious metrics that fall into this category are ad spend, return on ad spend, etc. In addition to these, businesses should also look at cost of goods sold (COGS), shipping expenses, and overhead. Many businesses forget to factor these costs when they look at what their allowable cost per acquisition can be.

This allows you to look at your return on ad spend through the lens of profitability, rather than just revenue generated.

What is the biggest problem businesses have with reporting?

Over-attribution. We see this issue with almost every client that we work with. Facebook, Google, and every other ad platform utilizes different attribution models. Generally, the platform will leverage an attribution model that favors them, and makes them look the best.

So Facebook utilizes an attribution window, meaning that if someone clicks on your ad, and then returns to your site within 28 days, they will claim that they produced 100% of the revenue from that client. Google defaults to last-touch attribution modeling, meaning that wherever that user came from when they made the purchase receives 100% credit for the revenue of that client. Other platforms count view-through conversions combined with an attribution window, meaning that if they saw your ad and then purchased within a certain time frame, that platform claims credit for that sale.

This scenario can lead to multiple ad platforms claiming that they are responsible for the exact same sale.

How do we combat this issue?

If you’re interested in learning more on this subject, we have a separate blog post on ways that every business can work through the over-attribution problem here.

In addition to those tips, our biggest suggestion for fixing this issue is getting a multi-source business intelligence tool. By extracting data from the back end of each of the ad systems, you can piece together a client’s journey and create your own attribution models. This allows you to see your true customer journey, rather than just a simple metric provided to you by a biased platform.

Unfortunately, even that solution relies heavily on the tracking that you have in place. If your tracking hasn’t been set up properly, then you have to rely on the data reported by these platforms, rather than leveraging and creating your own.

Your output is only as good as your inputs-

It doesn’t matter how much you spend on your powerful tools, they still rely on the data that you give to them. If your tracking isn’t set up properly, it’s impossible for a dashboard to correct that for you.

Powerful insights require great data. And unfortunately, good data requires great tracking.

Good news though, if you can get your tracking nailed down properly, then everything else glides into place. The old adage of “measure twice, cut once” applies to data as much as carpentry.

Going back to the metrics mapping process, we want to help you find the “source of truth” for every metric that you measure.

The “source of truth”-

Every metric should have a place where the definitive answer lives. If you want to know how much revenue you’ve brought in over the last month, you can check your bank account, or Stripe, or Paypal. If you want to know how long visitors from Instagram stay on your website, Google Analytics could help you find that answer.

Each data platform specializes in different data points, and we want to get the best data from the best sources.

Where to begin?

We recommend that every business start with the projects that will move the needle for the business. This generally means starting with sales and marketing initiatives, as they generate revenue for the rest of the business.

We have been shocked at how many issues businesses solve by getting their data set up properly for sales and marketing. Also, by leading with these departments, we can generally start to uncover holes in other parts of the business. If we see a spike in cancellations that coincides with customer survey emails, we know that we clearly have something to fix there.

It’s important to remember with every data initiative that it’s a journey. As much as we wish that we could fix every data problem overnight, it takes time to solve these issues and answer these questions. From there, we need to take action from the insights that we gained, and then we can see the results.

Leverage your uniqueness into growth-

As we stated earlier, every business suffers from terminal uniqueness. While this can complicate projects, that is also the place where you can see the greatest results.

By leveraging your uniqueness in the things that you track, you can get extremely granular, and explode your business in ways that others can’t.

One of our clients, Fancy Sprinkles, found amazing insights by tracking what no one else bothered to track.

Fancy Sprinkles, which does exactly what their name sounds like, gets most of their leads from Instagram. They decided to go back through every post that they had ever done and manually put into a spreadsheet the variables of the post. They tracked whether the product was shot indoors or outdoors, close or at a distance, color palettes, everything.

When they overlaid this data with their social media engagement rates over time, they found amazing insights. During October, they assumed that they should post something orange and black to capitalize on the holiday. But after consulting their data, they quickly realized that those colors got the worst engagement in October. The data told them that they should use purple and green, outdoors, and close-up. Their engagement skyrocketed because of these insights.

Tracking may require an up-front investment. Fancy Sprinkles needed interns to work for hours to catalog all of that data, but once they had the historical data, it became much easier to simply input those data points on every post that they made.

When should people seek help with their data?

As soon as it gets annoying or frustrating.

This may seem simplistic, but it doesn’t make sense for you to abandon your superpowers only to beat your head against a wall.

Your company hired you because of your skillset, and if data doesn’t fall into that skillset, it’s better to outsource that than to take away time and energy from the things that you do best.

What should businesses have in place before consulting with Praxis?

We built Praxis to meet companies wherever they are on the data maturity spectrum.

If you need help with your tracking, we offer Google Analytics audits and implementations. We even have courses that can walk you through setting up your tracking at your own pace.

If you already have your tracking in order and want to move on to scaling your business and gleaning better insights, then we offer pre-built dashboards that can help you start leveraging your data into growth.

If you have issues unique to your market or business that you need specific help on, we offer custom dashboards and implementations that we can build from scratch to better suit your needs.

Praxis Metrics- Data Predictions

Predicting data trends (2018-2020)

For this blog post, we wanted to revisit a guest episode that we did with the Vision Tech Team. This episode was filmed at the end of 2018, and focused on the changes that we saw in the data landscape in 2018, as well as predictions for 2019.

We found this episode particularly relevant given the changes in the data landscape that we detailed in last week’s post (which you can find here).

Enjoy this short podcast episode, and our insights below.

Insights from 2018:

  1. The data landscape changed rapidly, making it very difficult to predict what different platforms would do.
    1. Infusionsoft became much more open with their data in 2018, while Facebook began to clamp down on what data they would allow access to.
  2. Businesses began to understand the value of data, and started to understand it’s importance.
    1. Similar to the dot-com boom, businesses that fail to take advantage of their data started falling behind in the market.
  3. Data overload started coming to a head. Businesses started needing to figure out what data actually drives results.
    1. Trust in data became an issue. Data validation became a necessary part of data projects.
  4. Data visualization became the new buzzword.
    1. Everyone became obsessed with visualizing their data.
    2. This often came at the expense of actually driving new insights, as they would simply slap graphs on the same data.
  5. Automation started to spread
    1. In the beginning, nobody tracked anything.
    2. Then, businesses started tracking, but the data was stuck in silos.
    3. From there, data nerds started creating complex pivot tables and Excel sheets to bring the data together.
    4. Now, automation became feasible, and started taking over. This removed the need for manual reporting, and made the visualizations better.
  6. Going back to data overload, tracking exploded in 2018.
    1. You started being able to track anything that you wanted, and this lead many companies to overload on data.
    2. For those that actually leveraged this data, it caused explosive growth.
  7. Chatbots exploded onto the scene.
    1. Chatbots began to expand their footprint, with very promising results.
  8. Businesses began realizing the value of data democratization.
    1. Rather than having just one data nerd knowing everything about your data, businesses began to share data with all employees.
    2. By opening up data to the entire organization, you can gain insights that you previously would have missed out on.
    3. This allows organizations to leverage the combined brain-power of their entire team, rather than just the select few.

Predictions for 2019:

  1. Increased competition for data.
    1. As more companies realize the value of data, more and more companies will start competing for it.
    2. As data becomes more ubiquitous, companies will get better and better at targeting and marketing. This will drive inefficient organizations into the ground, leaving only the best to survive.
  2. Data overload to the max.
    1. As more data becomes available, businesses will need to start deciding what metrics actually matter to them.
    2. Most businesses will need to hone in on the handful of metrics that actually drive results (80/20).
      1. Those that capitalize on this principle will thrive in 2019
  3. Increased transparency and accountability.
    1. As more companies move towards data democratization, we will see a shift towards increased accountability across organizations. This will create cultures that thrive on extreme accountability.
    2. Since everyone has access to the data, it will become much more difficult for low performers to hide behind the work of others.

Looking back on 2018, things seem so much simpler. They only worried about compliance with GDPR. In 2019, we have a host of regulations on the horizon to worry about, decreased access to data and information, and platforms throttling data.

Looking forward to 2020, we’re likely to see an acceleration of the changes from 2019. We expect more data regulations to come into effect, and as a result of this, we expect a decrease in marketing efficiency.

One of the things that every business should start doing is gathering data into a data warehouse. This protects you from the whims of the larger platforms, and gives you complete ownership over your data. By gathering your knowledge and wisdom into your own database, you insulate yourself from the storm that is brewing on the horizon.

Praxis specializes in data, so we can help you take ownership over your data, create back-ups, and help you understand your data better than ever before. You can learn more about our products and offerings here. If you have immediate questions, you can contact a Praxis Metrics data expert directly.

Praxis Metrics- Chatbot KPI's

4 Fantastic KPI’s that you should be using in your chatbot marketing

How do I measure the effectiveness of my chatbot marketing?

We decided to have a “chat” with the king of Facebook marketing, Curt Maly, to get his thoughts.

Curt is a social marketing expert, owner of multiple online marketing businesses, consultant and national speaker joins us today with his vast knowledge of all things marketing data related.

What are the benefits of chatbots?

People generally react faster to the notifications from a Facebook message than to notifications about an email. Not every customer responds positively though, most people tend to the extremes in their feelings towards these messages: they either love them or they hate them.

Another benefit of using chatbots in marketing is that they help to automate out customer service. Chatbots essentially create auto-responders within an AI platform, allowing you to dedicate your time and resources elsewhere.

Chatbots also help decrease the necessity of traditional funnels. You can utilize chatbots to move your clients from one point of the traditional funnel to the next without needing to set up traditional stages. Rather than segmenting people into phases like a traditional funnel, chatbots allow you to just have a conversation with the client and naturally progress them through their journey.

How hard is it to set up chatbots?

Just like funnels, you can make chatbots as simple or as complex as you would like. Creating a basic chatbot requires minimal knowledge, and only needs a few options for auto-response. Essentially you just need to think through a standard conversation that happens on a sales call or customer service call, and input the different variables into the chatbot.

KPI #1- Cost per Acquisition

This is the amount of money that it takes for someone to engage with your message. You can find this information by dividing your ad spend across the total number of new messages that you get. Facebook will try to conflate this data with Cost per Reply, but that will give you an inaccurate picture of your actual cost per acquisition.

KPI #2- Cost per Reply

Facebook will track this metric automatically for you. It’s a very important metric to keep top of mind, because it can help you understand the cost of re-engaging a lead that you fell out of contact with.

KPI #3- Cost per Open

This metric matches up nicely to open rates on email, with one difference: most people see at least an 80% “open rate” on Facebook messages. This happens because if someone browses Facebook on their browser, the message will automatically open, inflating the open rate.

KPI #4- Click Through Rate

Click through rates on Facebook messenger generally align very closely with email click through rates. Most people will see between 6-10% click through rates on Facebook messages. Many people mistakenly claim that they get better click through rates through Facebook messenger. In reality, they are generally getting more leads than from email, due to the increased open rate of Facebook messages.