Category: Technical

8 Essential Database Performance Metrics You Need to Track
In general, databases play a vital role in any information system. It can store, update, and query essential business data. That’s why a database system’s availability, performance, and security are critical concerns for any database administrator.

To facilitate this, you need to ensure that you’re using different database monitoring tools.

Why Do You Need to Monitor?

When you monitor databases, you’re doing it just like you would monitor every other component of your IT system. It keeps you well informed. That way, you can make better decision-making in the future.

Databases are the dynamic indicator of the health and behavior of a system. Its impressive performance directs you to particular problem areas. Once you identify these issues, you can then use these metrics to help you with the debugging process.

But databases have a sea of metrics that can overwhelm anyone. So, it all comes down to the metrics you need to monitor. One of the most straightforward approaches is starting with the basics.

In this post, we’ll talk about the essential database performance metrics that you need to track:

1. Response time

Data response is vital to your business’ performance metrics. This is the average response time for every query on your database server. There are several ways that you can approach response time. First, you need to assess the live database response as a solo number on a dashboard which you can share with other developers in your system.

You can use a column graph or a line chart to measure the average response time for the query that your database server gets.

But what if your database shows a more significant average response time? Then, the best solution for this is to optimize your queries and enhance concurrency via program analysis.

As a developer, it’s part of your job to prepare yourself to handle concurrency problems regarding data democratization for users. This approach is practical for people that go through data within their dashboards using various business tools and mechanisms.

2. Availability and resource consumption

This is where you must regularly keep up with databases, ensuring that they’re online and function regularly.

These tests should be done in both peak and non-peak hours. It would be best if you didn’t have to scan since your monitoring will catch any changes manually.

3. Database throughput

Data throughput is one of the most critical database performance metrics out there. It is the volume of work that’s done by your database server over a unit of time. Or it could be per second or hour. However, you can usually measure this as the number of queries executed per second.

It lets you monitor how quickly your server processes any incoming queries. If it’s less than the number of incoming queries, then this could cause your server to overload. As a result, you’ll have longer waiting times for each query, which in turn, slows down your website or application.

If this is the case, then you might need to upgrade your server infrastructure, or you can optimize your queries. That’s why having good database management support is essential.

4. Number of errors

Your database will return an error response code if a query doesn’t run successfully. You need to track the number of queries you get for every error code.

That way, you can easily find out which errors frequently happen and how to fix them.

5. Query performance

What’s the reason for a substandard query performance? Well, the answer lies in the query. You can look for issues in queries that often take too long. You can find necessary data and retrieve time and other important information. The objective here is to track slow queries that tend to hurt performance.

Let’s say you’re running an enterprise app with a MySQL server. Your database queries when a user logs in, assessing the login credentials.

Slower queries hurt how your database performs. This specific query also significantly impacts an enterprise app’s overall performance if users refer to it frequently.

But what if your resources aren’t overwhelmed, and you still get poor database performance? Well, more often than not, there are several reasons behind this. It could be because of inefficient query plans, unmanaged database statistics, and schema modifications.

If you want to troubleshoot such issues, you have to be equipped with knowledge related to database internals. It means you’re referring to filters, query plans, and other things your database query optimizer uses.

6. Scheduled tasks

A database often runs repetitive tasks scheduled as “jobs.” Systems like Oracle and Microsoft SQL Server have these built-in scheduling facilities. Meanwhile, others have third-party schedulers.

Some examples of these scheduled jobs include:
- A full and incremental database backup
- Applications-specific tasks such as nightly data loads, archiving, and exports.
- Database maintenance like analyzing, updating stats, database integrity checks, compaction, log rotation, reindexing, and more
No matter the function, the outcomes should be monitored, whether they succeed or fail.

7. Buffer pool usage

Usually, a buffer pool or cache uses as much memory you need to allocate to it so that it will accommodate maximum pages of data.

When the pool fills up, it will eliminate older or less used data, making room for newer information. It’s also possible for you to access information related to the buffer pool. It lets you know everything about the data the SQL servers store in the memory.

With a dynamic management view, you can identify every row for each buffer descriptor. This looks for and gives you relevant information about every memory page.

However, this may take time to query this on a server with extensive databases.

8. Most frequent queries

It’s also vital for you to track the 5-10 most frequent queries from your database server, along with the average latency and frequency.

You can experience a significant performance boost in your database by optimizing these queries.

Over to You

So, there you have it. Database performance metrics are there because it helps you with database performance monitoring. You also need it if you want to optimize your business.

Hopefully, the following database performance metrics we’ve mentioned will enhance the performance of your database.
August 17, 2022
Understanding Your Google PageSpeed Score

You’ve got a shiny new website all ready to showcase your store’s products. You’ve invested in great product photos. You’re ready to slip the switch and make some money.

So, you launch, investing your first few hundred dollars or so into some targeted Facebook ads.

Crickets. No conversions. No sales.

You look at your stats and, while the ad got good engagement, most of those clicks didn’t translate into full-fledged page views.

Especially on mobile.

“What happened?” you ask yourself.

We explain what happened.

PageSpeed matters

Have you ever clicked on an ad, interested to learn some information about a cool new product, service, or company, only for the website to take forever to load?

What do you do?

Easy, bounce.

You have better things to do than wait several seconds for a page you were only nominally interested in to load.

So, you leave.

The same thing happens to you, when your website is slow. And slow websites are expensive — especially for ecommerce companies — because conversion rates drop by xx as page load time increases by 1 second.

For an ecommerce company, conversion rates and revenue are closely linked to one another.

Testing your website’s PageSpeed

Before optimizing your website for speed, you have to get a baseline.

First, head over to Google’s free PageSpeed Insights Tool. Enter in your website’s name (ex. www.mywebsite.com) and click “Analyze”.

You should get something back that looks like this:

Credit: PageSpeed Insights

Right away, you’ll see a score for this page, based on several factors, discussed in more detail below. As you can tell from the legend underneath the score, a score of 0-49 means your page speed is poor. A score between 50 and 89 is considered competitive. A score above 89 is considered good.

By default, this test only analyzes one page, on a mobile device. More specifically, a mobile device on a 3G network. The idea is simple: if your page loads fast under those conditions, it should load fast under much better conditions, such as 4G networks, fiber connections, or desktop browsers.

Because the test only scans one page at a time, you should test a few different pages to get a sense for overall performance. Or request a complete site analysis from your website developer.

It’s highly likely that pages like your homepage — which may have product sliders, social media feeds, or the like — will run more slowly than pages like your blog, which are largely text. So, run the test a few times on a few pages.

Understanding your score

Once the test runs, you’ll see a section that looks like this underneath your overall score.

Credit: PageSpeed Insights

These metrics can help you start to get a sense for where/how to start improving page speed scores.

But, first, you should know what they mean.

First Contentful Paint

This refers to what marketers call ‘above the fold’ content — the first thing(s) you see when a website loads. For most people, it’s the banner image & text on your homepage. This metric matters because until this First Contentful Paint occurs, the person on your page doesn’t have anything to look at. During this time, they are left wondering what’s going on on the website and are highly likely to bounce if this exceeds 1-2 seconds.

Time to Interactive

Just because you can see a page, doesn’t mean you can do anything on it. Imagine going to google.com and having to wait 30 seconds before you could use the search bar.

That would be terrible and you probably wouldn’t use Google anymore.

Same thing with your website.

While First Contentful Paint shows how long it takes for a user to see something, Time to Interactive measures how long until the user can do something.

Total Blocking Time

Webpages are made up of a lot of code that the browser deciphers into images and text we can see.

But computers don’t see things like we do.

They parse the code.

Top to bottom, until it’s all parsed.

Except sometimes there are scripts or code snippets that have directions to stop parsing until they finish executing. For example, you might use custom fonts and you (or your developer) set up the website so that it doesn’t load until those fonts are available, so it always ‘looks right’.

When a browser can’t parse the content it wants to parse because it’s waiting for some other code to finish running, this is called blocking. The code that hasn’t finished running is blocking all the code behind it from running.

And it’s not a good look.

Largest Contentful Paint

While First Contentful Paint measures how long it takes for the first few elements to be visible to a user, Largest Contentful Paint measures how long until the lion’s share of the content is visible. Once the Largest Contentful Paint has loaded, the page is generally considered “done loading” and the browser is ‘done’ with most of its work — meaning the user should now be able to consume the contents of the webpage at their leisure.

Speed Index

The Speed Index describes (about) how long it takes for the full page to load.

Cumulative Layout Shift

Have you ever been on a website like Forbes when it shifts up or down, so that an ad can be served?

Super annoying, right?

Cumulative Layout Shift measures the degree to which elements on a page move after it’s loaded. The lower the better because users don’t want their experience interrupted unpredictably or without their express consent (e.,g clicking a button).

How to Improve PageSpeed

Knowing a little bit about how the web works and what the different sections above mean is a great way to start improving your PageSpeed scores, increasing conversion rates, and selling more products.

In general, you’ll see a section on your report that looks like this:

Credit: PageSpeed Insights

Unfortunately, everyone’s report is going to look different, so it’s a bit hard to offer specific advice to you without an assessment of your website, but there are a few culprits we see that, when addressed, can radically help improve PageSpeed.

Manage images properly

People are visual and so it’s natural to want to have images on your websites. Fortunately for us, our hi-def cameras and phones take beautiful images.

Unfortunately, though, these images are often massive and bloated, which means they can take a long time to load. Large images at the top of the page will make both your First Contentful Paint time and Largest Contentful Paint times go up.

Your best bet, when uploading images to your ecommerce store, is to shrink or crop them before uploading them to your store. Generally, you don’t need images wider than 1000px (except maybe for an intro slider), so resize images down small enough to increase speed, but large enough they aren’t grainy.

If possible, use .jpgs for photos and .pngs more sparingly. JPG files are usually smaller file sizes and don’t need to be as ‘crisp’ as .png images (like logos and icons) need to be. If you have the technical skills, use .webp images — which are both small and hi-resolution — but have a plan for supporting browsers that don’t support .webp (like Safari).

Finally, ‘lazy load’ images towards the bottom of the page. Lazy loading refers to the practice of not fully loading a resource (image) until the user gets to that part of the page. That prevents the image from being a “Blocking” resource as the page loads.

Clean up your Javascript

The second thing we see a lot of are code snippets in the ‘head’ tag that increase the First Contentful Paint and Total Blocking Time metrics.

When a browser gets to a 3rd party script, it very often has to go off-site to where that script is stored, process it, and then come back before it goes on to the next line of code.

While this might only be a few hundred milliseconds, those milliseconds add up to real seconds. And can destroy your PageSpeed scores.

If your First Contentful Paint and Total Blocking time are greater than 2 seconds and 1 second respectively, this area merits a look.

Remove/reduce unnecessary code

While being able to add and edit your own website content is really helpful, it can also make getting a fast website really difficult. Page builders and Content Management Systems’ (like WordPress or Shopify) themes can bloat your code.

You can easily end up with stylesheets, javascript, and 3rd party resources being loaded on pages where they are not needed.

The result is a site that is unnecessarily slow, but ‘easy to use’.

Fortunately, there are tools your developer has that can allow them to load scripts on certain pages, while excluding them from others. It might take a few days, but they can probably help you load scripts smarter.

Small tasks, big wins

When working on improving PageSpeed, it’s very common to see poor scores across the board. Part of this is the cascading nature of the report — a long First Contentful Paint will generally cascade to a long Large Contentful Paint (afterall, if the first part hasn’t loaded, how can the largest part load?) and long Speed Index. If code bloat is a problem, Time to Blocking and all the Contentful Paint metrics will be high.

So, start with factors related to your First Contentful Paint, such as reconfiguring (or eliminating) Javascript and code bloat. These small tasks will cascade through the rest of your metrics.

Next, address your images, since it’s manageable without a developer — albeit laborious.

Finally, have your developer review the remaining items and work with you to come up with a plan to handle the rest.

By the time you get there, though, you will probably have realized substantial benefits in PageSpeed and should start seeing your conversion rates and sales increase.

Nobody likes a slow website and, for ecommerce companies, a slow website impacts negatively on sales. Head over to Google’s PageSpeed Insight tool, evaluate your website, and get started trying to optimize for speed today.

October 2, 2020
The Link Between Ecommerce Customer Relationships and Data Analytics

Customer relationships are as important in ecommerce as they’ve ever been in traditional retail — perhaps more so, because online merchants are competing in a global battlefield, meaning that providing good store experiences really matters to companies operating in this space.

Having a brick-and-mortar store in the right location with the right aesthetic can be enough to pick up some easy sales, but the closest online equivalent to having a good location is ranking well for prime keywords, and that’s much harder to achieve.

Data analytics, meanwhile, fuel the engine for large-scale growth. The accessibility of rich data on every trackable element of online activity (and that’s a large pool) is what allows businesses to comprehensively optimize their operations. Using it to full effect, today’s big brands are persistently finding new and smarter ways to improve efficiency.

So what’s the connection between these two mainstays of the modern merchant? They seem disparate — one emotive, nuanced, seemingly analog, and the other objective, exact, firmly digital — but they inform and contextualize each other. Follow along as I provide a more in-depth look at what I mean by this:

The analog/digital divide is illusory

When machine learning first started getting some traction in mainstream business, it met with some resistance from old-fashioned types who didn’t see its value for complex processes like customer relationship management. After all, they reasoned, people are difficult enough for other people to understand — they’re capricious, awkward, often ignorant, and unpredictable.

On the other end of the spectrum, those who were firmly on the digital bandwagon might have considered the investigation of specific customer cases to be unnecessary. When you have pulped and parsed data from hundreds or even thousands of customers, what’s the point in getting hung up on specific events? That’s how you miss the wood for the trees.

What smart retailers eventually realized, though, is that the inferred distinction is false. It isn’t a matter of analog versus digital — it’s a matter of micro versus macro, and if you want to make the most of your online sales opportunities, you need to consider both. Here’s why:

Performance data must be contextualized

Let’s say (as is entirely plausible) that you either run or work for an ecommerce business. Your system is configured to gather information through Google Analytics or whichever analytics platform you prefer, so your database continues to grow. You thus have access to countless metrics of varying significance: everything from dwell time to average conversion value.

You can spend hours poring over that data, but that time won’t necessarily produce any notable conclusions. The problem? A lack of context. When you look at performance metrics, you see only fragmented reflections of complicated situations. You see how many pages the average visitor looks at, but you don’t know why, so you can’t know what (if anything) it means.

The value of an involved customer relationship management process, then, is in giving you the framing you need to make sense of your data. The more feedback you gather from your customers, the more you can read into your metrics — whether you’re able to understand why a certain page is performing so well, or left with a curious disparity between what your customers say and what your data says (this can be remarkably illuminating, because people are often mistaken when trying to detail their own behavior).

You can’t usefully pursue CRM without analytics

On the other side of the coin, of course, you have the prospect of dedicated customer relationship managers trying to succeed using little more than direct interaction, buyer personas, and nebulous concepts such as gut feeling. It’s extremely valuable to know how to interact with people, yes, but it simply isn’t enough if you’re trying to achieve optimal results.

As noted, we’re not that great at explaining (or even remembering) our own actions. Think about your favorite websites, ecommerce or otherwise: why do you like them so much? Are you absolutely confident that the reasons you can cite fully explain your attachment? Do you recall precisely why you clicked on the link to this piece? However certain you are, I’m inclined to suggest a sliver of doubt. Our subconscious minds heavily shape our decisions.

But the issues with self-assessment don’t stop there. Consider that the human race is excellent at self-deceptions, be they harmful or innocuous, and a great example is that of musical taste. In principle, we should all just accept that we like whatever music we like, and feel no shame or guilt about it — but that’s not how culture works.

Here’s an specific scenario: let’s say that you regularly browse and buy from a music store (whether involving physical media or digital downloads), and one day someone from the store reaches out to you for some feedback about your store experiences. Because you’re embarrassed about your musical taste and you feel confronted, you offer a warped account of the features you like and dislike, resulting in feedback that won’t be too useful.

Our retail habits can be deeply personal and emotional, so our recollections are invariably sullied by ambiguity, redaction, and creative reimagining. But analytics data doesn’t degrade, and while it can be unclear, it can’t be mistaken. Only by using it to underpin your CRM can you establish the foundation you need to achieve consistent results.

The future is driven by integration

In some ways, you can draw a parallel between this link and the one that binds all marketing channels together in modern multi-channel marketing. Retailers that would once maintain distinct departments for distinct tactics — a social media team, an email marketing team, a paid advertising team — now understand that they’re better understood as parts of much broader customer journeys.

Similarly, it isn’t ideal to have a team concentrating on CRM and another handling data analytics. Data analytics should be the connecting tissue between all departments as opposed to a pursuit in itself, because context-free data holds little value, but contextual data is key to turning a small online store into an international retail powerhouse.

Aiming to get more from your ecommerce store, then? Start concentrating more strongly on weaving analytics through everything you do. It’s become highly accessible in recent years, particularly for those using high-end solutions such as Magento’s Commerce Cloud edition or Shopify’s multi-channel centric Shopify Plus — especially since each one (along with various other retail platforms) has straightforward API connectivity with Adverity DataTap. So what are you waiting for? Stop treating data as a gimmick, and start getting results.

Overall, then, the link between ecommerce customer relationships and data analytics is simple: it’s codependency. You can’t have valuable ecommerce data without understanding the customers that produce it, and you can’t manage customer relationships effectively with no way to reliably track the consequences.

July 18, 2019
How AI Is Changing Ecommerce Web Development

Succeeding in online retail isn’t easy, but entering it might be. The barriers to entry are already low, and now AI is entering the equation to speed things along. Here’s how:

Ecommerce is among the most adventurous industries in the digital space. With so much money changing hands, and even the simplest virtual store capable of selling worldwide, it’s easy to understand why even the biggest brands are willing to try new things. If something doesn’t work out, no big loss. If it proves profitable, you can scale it up right away.

This is of particular interest when it comes to AI, something that has already proven somewhat impactful across most industries. Today, it’s starting to change the nature of ecommerce web development at a fundamental level. Let’s take a look at what this actually involves:

What AI Means in a Practical Context

AI is a term that is often used in a vague or unclear way, and this produces a lot of confusion about what it actually entails. As such, it first bears noting what AI means here, and that’s the use of technology to mimic (and thus replace) flexible human behaviors. AI software differs from programmable routines that simply go from start to finish because it is able to adapt (to some extent) to its surroundings.

The key takeaway here is that we’re not talking about a like-for-like replacement for the human brain. AI is a long way off that type of achievement, and it may never reach it. Even if you could trust every part of your online operation to an AI suite, it wouldn’t be a good idea in any sense. Today, at least, AI is an advanced tool that can complement your everyday work instead of taking it over and getting it done for you.

That’s how it fits into the world of ecommerce web development. As we’ll see shortly, it’s starting to work with developers and retailers to get things done more effectively and efficiently.

Its Role in Automating Store Creation

Particularly at scale, ecommerce can take a lot of work to carry out. Products must be updated, stock managed, orders processed, and support requests routed — and that’s once you have a store up and running. In principle, the development stage may be even more frustrating. The more your store will be offering, the better it needs to be. There’s certainly no shortage of competitors out there to take your business if you provide a weak UX.

But this process of starting from nothing and building up to a fully-fledged store isn’t the chore that it was in the early days of the internet. The establishment of the modern ecommerce CMS with its setup wizard and automated suggestions turned the online store from a costly bespoke project to a factory-line assembly viable for rapid production and business trading. Instead of typing code, you can use a GUI to assemble the necessary components (with the code drawn from existing libraries) — often taking a WYSIWYG approach for ease of use.

So where can AI enter into the equation? Think about how much of the development process must currently be planned out carefully, then consider the purpose of theme templates — instead of creating a theme from scratch, someone eager to get on with things can find a theme they like and mostly repurpose it. AI can use mix-and-match elements to build entire stores based on answers to some configuration questions (e.g. “What products will you be selling?” or “What’s your unique selling point?”).

At some point, you might even be able to say “Create a luxury cosmetics store” and have a store builder use your internet history and associated profile to create something to suit your tastes. There will always be room for manual intervention, naturally, but it would be a superb way to get a strong baseline very quickly and with minimal effort.

How it Benefits from the Integrated Cloud Model

The reason that we can realistically envision that type of AI solution is that ecommerce has all the required resources. Multi-channel stores search for customers in every corner of the internet, the whole process enabled through smart integrations, and vast quantities of rich data are created. When a customer signs in to a store through their Google login, it syncs yet more data that can be used by AI applications to develop (or adjust) stores.

Imagine a future of leaving an AI admin in charge of your store’s design (and even its offers). An AI could consistently review competing content from comparable sites before making adjustments accordingly. Is the color blue proving popular? It’s easy enough to adjust hues. Is a new resolution of screen becoming standard? Resample images to better suit that size and aspect ratio.

At the same time, it could monitor internal analytics and adjust placement to run smart A/B tests instead of requiring them to be configured manually, and even use NLP to generate customer surveys to gather invaluable feedback and parse the results. It would need some oversight, naturally, but at some point it would mostly be able to operate independently.

When the modern easy-creation ecommerce CMS became mainstream, a lot of the arduous work behind building a store was automated, but plenty of tasks were left to be handled manually because the technology wasn’t there to deal with them. With so much data available online now, and AI software becoming increasingly powerful, we’re now starting to see more and more flexible functions taken over — leaving store owners with yet more time to focus on selling.

May 17, 2019
How To Select The Best Hosting For Your Ecommerce Store
It’s no use having a fantastic ecommerce store if the hosting isn’t up to par. You might as well work on an impressive indoor art installation and then lock the doors and shutter the windows so no one can actually see it. Your store host makes sure that everything works as it should — that people can reach your store and order through it without experiencing any major delays.

But it isn’t as simple as choosing from a set of web hosts ranked on a simple linear scale, because each has its advantages and disadvantages. You must think carefully about what your particular store needs: what platform it uses, what speed it demands, and what level of support you’ll likely need from a host.

Here are some tips for choosing the optimal host for your ecommerce store:

Find a hosting provider that suits your store’s platform

While many hosting providers — most, in fact — will gladly (and competently) host a website using any functional platform, it’s common for any given hosting provider to adopt a strong focus on one specific platform. The practical consequences of this vary, but if a host mentions a preference for a certain platform, you can reasonably infer several things:
- They’re familiar with its unique challenges.
- They can make the setup process straightforward.
- They have some understanding of how the platform is updated.
- They can meaningfully predict its resource demands.
This is important because you’re always going to encounter some issues with your CMS of choice, no matter how polished it may be (and no matter how many resources its creator can bring to bear). When that happens, you can of course try to resolve them yourself, or seek support from the wider online community — but being able to go directly to your web host will save you a lot of time and effort.

For example, SiteGround is a very popular host that places a lot of emphasis on its use for WordPress (and WooCommerce) websites. In fact, it has committed so strongly to that angle that WordPress.org actually recommends it specifically. If you’re interested in using WooCommerce for your store, then, you’d get excellent results with SiteGround.

Read plenty of varied reviews

Critical reviews from expert sites aren’t likely to prove sufficient when you’re trying to get an idea of what you can expect from using a hosting service. They’re exceptionally useful in general, of course — consider the nature of this site! — but they don’t tell the whole story.

In many cases, for instance, a hosting service will offer a critic an extended trial of a service, and have the opportunity to allocate extra resources to ensure that everything performs optimally. This leads to the resulting review reflecting an experience that has little in common with the service provided to the average user.

Similarly, consumer reviews aren’t always particularly useful. Conventional users can’t be relied upon to do much research before they offer their thoughts — an unsatisfied customer might bemoan the lack of a particular feature that’s actually present in the system (having simply missed it), or slate the company for its pricing despite comparable services being no cheaper.

What’s truly important is that you form your own impression in the end: one that’s as well-rounded as possible. Only then will you be able to make a fully-informed decision about which hosting provider to choose.

Carefully consider support packages

You may find a hosting provider that boasts a great amount of relevant expertise, but simply having that expertise won’t do much if they prove unable (or unwilling) to share it with you. And when things go wrong, you need to know that you can get some assistance in a timely fashion. Savvy advice that only arrives a week after your site stopped working won’t help much.

Very often, beginners to ecommerce don’t give too much thought to support requests, assuming that they’ll be able to figure things out as they go. This is a mistake. The quality, nature and convenience of your hosting support will prove vital. Once you’ve established a solid shortlist of viable contenders, narrow them down using support as the primary factor.

At a minimum, consider each of the following:
- Their active support hours.
- How quickly they can respond to requests.
- The channels they provide (e.g. email, telephone, live chat).
- How they come across in conversation.
- What limitations (if any) they place on their support services.
You might find that a support package that sounds good actually has small print stating that the company can decide to stop providing support if it feels that you’ve placed too many requests. Do you want to risk needing assistance during a time of crisis but seeing a fair-use policy cited as a justification for leaving you in the lurch?

Wrapping up, following this process — lining up hosts that cater to your platform, filtering them using reviews, then filtering them down to optimal candidates based on support provision — should leave you in a strong position to decide. Commit time to the matter now so you can proceed with confidence that your store will prove robust.
March 11, 2019