A Systematic Approach to Improving the Efficiency of an Existing Web Project
There are many different materials on the Internet about how to properly create new Internet projects, how to promote them, how to improve an existing interface, etc. But, there’s no systematic viewpoint on how to support and develop existing projects. Many people think that it’s like buying a nice new car: so, you’ve purchased it and sometimes you change the oil, or you may decide to buy a new car rug in a couple of years. In fact, this isn’t a good approach: it’s not enough to pay for hosting and the domain once a year. Sometimes, it’s not even enough to make improvements. A website is a work tool, and it should operate as efficiently as possible. All successful websites demand continuous and uninterrupted attention.
In this article I will try to systematically look at the problems related to existing websites, and offer recommendations on how to constantly increase their effectiveness. This material will be useful to owners of existing projects, and will allow readers to comprehensively consider issues related to support and development with reference to economic efficiency, and for some, perhaps, it will breathe new life into an old project.
What Usually Happens?
An entrepreneur conceives a new Internet project. It’s created by a contracting company or a team of employees. A promotion campaign is launched. Then, if the project idea and implementation are qualitative, the first users appear and the first revenues arrive. Under a positive scenario, the project at some point exceeds the break-even point and often turns into a cash cow. This is especially true for projects in the field of e-commerce. It’s at this point that support and development for the website often slip into sluggish editing of bugs and, perhaps, a design renewal every five years.
More often than not, nobody monitors the effectiveness of the website. Well, perhaps if a marketing manager works on the project and sometimes looks at Google Analytics. Some even have marketing reports that show the number of visitors. Need more sales? Let’s increase the budget for contextual advertising or expand the semantic core for SEO. Not many people consider commonplace tools, such as website goals and conversion tracking, never mind calculating the cost of a visitor or analyzing effectiveness through channels. This happens because everyone “writes code” in outsourcing and no one knows how to work with products, while entrepreneurs themselves don’t really understand what to do with their websites.
No one follows the interface. It would be very interesting to see how many website owners use a click-through rate. I think that most of them have never used it, whereas it’s a basic tool that allows you to monitor interface effectiveness. Very often, the interface is created and forgotten until the website has been completely redesigned. Few people remember the large projects that repeatedly conduct A/B testing in order to analyze and improve interface effectiveness. It’s namely the interface that determines conversion, and hence the effectiveness of all marketing efforts.
You can achieve sales growth in several ways: for example, by increasing the marketing budget, which will result in more website traffic, or by improving the interface, which, with an equal number of website users, will result in more sales. The second option is much better and should be at the top of the to-do list, that is, we must first achieve maximum conversion, and only then can we invest in advertising so that every dollar is spent with maximum return. Or, they should be implemented together at least.
Technological issues are also complicated. The technology on which the website was built gradually becomes outdated. New language, framework and CMS versions appear. No one wants to update to the newest version as it doesn’t make much difference in terms of business. What for? After all, the website works properly! A lot of the old code accumulates in the project because there’s not enough time or money to work on code refactoring during development. No one has written any unit tests, and errors accumulate. Support costs increase continuously.
The technology sometimes becomes so outdated that even official documentation is hard to find, and the abundance of outdated code seriously hampers the programmer. And yes, programmers often come and go during the project, and those departing very rarely leave any documentation. But, nobody notices it because the tasks have finally been solved, the project brings in revenues, and no one even thinks of considering the cost of support…
Everything works on the outside, but, in fact, budget funds are repeatedly and inefficiently wasted. The worst thing is that this is quite problematic to assess objectively, and developers aren’t ready to admit it. A week ago, I came across a particular case — a current client’s contractor specifically failed to update the framework version in order to work longer and get more money for the hours spent.
A familiar situation? Do you recognize yourself or any of your friends?
Primary Documents and Information
First, we need to make sure that we have all the required documents and access in order to continue our work.We bring the following up to date or create from scratch:
- Access to the site. First of all, via FTP and SSH. This is required for further development. Secondly, to the site management system.
- Documentation. Each developer must manage documentation: what he or she writes, where it’s stored, how it works, etc. The larger the project, the more documentation is needed. This point is usually neglected, and a new developer has difficulty understanding the old code. I think that documentation is available in less than 10% of the projects. If you don’t have it now, then it’s time to work on it, and then give it to another company for verification.
- Third-party services. Access to Google Analytics and Google Search Console, Webmaster tools. Some resources have a number of specific integrations; we must also have access to them.
- Marketing statistics. If you conducted advertising campaigns earlier, you should have statistics on them. SEO? — What words were used, how positions have changed, were there any filters, etc. PPC? — What words were used, what rates, what ads, etc. SMM? — Access to fan pages, brand mention statistics, traffic statistics, etc. The older the statistics that you have, the more effective your future marketing activities will be.
- Audits. If audits were conducted previously, the same things are required: there should be reports and results.
- Project description. Well, that’s when we have something understandable and standard, such as an online store. But, there are often very large systems that should be thoroughly described in a special structured document. For example, let’s take any rating system. A new developer will have a bunch of questions: “What formula is used for calculations?”, “What’s it connected with?”, “What’s it needed for?” etc. This is a common function that can be expressed by only one number on the interface, but it must be connected with everything on the site.
The larger the project, the more it changes over time. A good specialist, whether he/she’s a marketer or a developer, will start with questions about the past site and the current situation, and only then will he or she draw up recommendations.
I like to plan everything. It allows us to understand where we are now and where we’re heading. Website development is also a strategy where there should be coherent goals and clear KPIs. This is the only way to achieve efficiency. The strategy should disclose plans for the development of the site itself, and for its promotion on the market.
To begin with, we should listen to the users. We can do this by studying their behaviour on the site using Google Analytics, or other tools. We can directly ask for feedback through feedback forms or interviews. We can use the option of “reasonable borrowing” from competitors… and hundreds of other ways. Once again, it very much depends on the project. If it’s more or less standard, and there are many analogs on the market, Best Practices should be known; they just have to be applied. If it’s unique, then it’s harder, so you’ll have to experiment.
There are many different approaches to developing strategies; this isn’t about them. Resources are always limited for everyone, so it’s unlikely to do everything at once. I believe that we should start with the most important point, that is, with what directly influences profit. In particular, the development of the business component (everything related to customer development), optimization of marketing costs (it’s not about cutting costs, but about improving their efficiency), the interface (conversion and usability), and the quality of the technical solution (it should work stably and the support costs should be optimal). We must make analyses on all these fronts, and then start working out an action plan.
The algorithm for systemic work on the project is as follows:
- Carry out a comprehensive audit of the project
- Request user feedback
- Analyze direct and indirect competitors
- Identify the most anticipated and profitable functions for the project
- Streamline implementation priorities
- Work through each introduced item
- Track implementation efficiency
ROI should be considered for all these actions. There are things that should be done right away. For example, updating a technical platform can significantly reduce support costs, and although this update will require financing, it will pay off fairly quickly, that is, ROI will be high. Moreover, there are improvements that aren’t very profitable.
Large companies plan several years in advance. When it comes to an Internet project, even a small one, you need to plan at least a year in advance, and preferably three to five years in advance. Next, by constantly listening to customers, you can correct this strategy. Of course, everything is always changing on the Internet, but unless you have a goal, you’ll never achieve it.
This is what few people know how to do, but what the whole business generally depends on. This means developing the product around the client and his/her problems. The methodology is aimed at carefully and regularly analyzing customer needs and improving the product. This means working on things that bring in money. Furthermore, this methodology can be applied to all projects. The main thing is to answer the question: “What else does the client need?”
Imagine that we have a project that provides some kind of service or sells a certain product. How can we increase sales? There are thousands of ways: you can offer something for free with the product, you can improve delivery (speed, quality, service), and you can even come up with a new product as part of the existing project. Another example, closer to our reality: so far, many online stores haven’t extended credit for purchases. But, more than 50% of all goods are sold on credit offline! That is, many online stores lose a very large number of customers, although they may be quite successful, can sell a lot of goods, and don’t even suspect that they’ve lost a large percent of income. There are many more similar examples; they’re in every Internet project.
Are you really doing everything you can to develop your customer base? Or maybe they can purchase products more and more often? Few people ask themselves these questions, so huge profits are often lost.
Interface and Design
Now, the most important thing is to understand that a beautiful image doesn’t mean a convenient interface. Each site and each page has a goal(s). The effectiveness of the interface can be measured only by whether it reaches its goals or not. This doesn’t necessarily mean that something must be sold; it can be a registration, the number of pages viewed, writing a comment, etc. All these goals are measurable, and we need to work on improving each and every one of them. Old projects are problematic because such goals are often not even set. Many people think in very broad categories: “We have a product, so we must sell it.”
First, you need to define the main KPIs for the site and its most important sections. It’s better to do this with a UX / UI designer, who is well versed in the design and theme of the project. Next, you need to analyze Google Analytics, and study user behaviour. Then, you have to gather feedback from users and think about Product Development. Only then can you start improving the interface with all the necessary materials on hand.
Changes can be implemented in two ways:
- Blindly, when we rely on the designer’s experience;
- Through A/B testing, when the effectiveness of each experiment is measured.
There is no ideal option. In the first case, we have a high risk due to the human factor. Even the most experienced designer can make mistakes (we’re all human!). However, there are usually a lot of obvious things whose effectiveness has long been proven. In the second case, the test might be carried out incorrectly, and it’s not at all about questioning the knowledge of the specialists who conduct it. The small part of the site chosen for the test may have been too small; a lot of new visitors may have arrived on the test days; users may be used to the old interface, and the new one, even though it’s much better, just confuses them. There can be a lot of options here, so you have to be careful with implementation, but still it’s better to see changes in results than nothing at all.
The old interface can be regularly re-developed and improved. But, there are always new ideas and new developments for what we have now. It must be a continuous process, continuous improvement. Typically, the initial vision of the project, when its MVP is being developed, changes quite rapidly over time. This happens several times: first, when we begin developing it, and then, when the project is ongoing and the first user feedback reports start coming in. Sometimes, projects change not only externally, but they also change their main idea. This is another reason why you need to listen to users, and implement improvements that are based on their feedback.
However, many projects have problems with adaptability for different devices. Not only are many old sites not adaptive from the beginning, but even the adaptive ones usually have many problems and shortcomings. After all, there are so many different types of devices, and it’s not easy to take all of them into account. Adaptive design can always be completed or existing adaptive design can be improved, but, unfortunately, not everyone thinks about doing this.
In addition to designing the interface, it’s possible and necessary to work with layout and graphics. It can be optimised to make the site load faster. Users will be satisfied and search engines will display the site first in search results. You can check the site’s download speed here: https://developers.google.com/speed/pagespeed/insights/?hl=en.
You can check the correctness of the display in different browsers here: http://browsershots.org/.You can check for errors in the layout here: https://validator.w3.org/ — Google, which, by the way, not so long ago began to take validation errors into account, and the more errors there are, the lower in the page ranking is the site.
So, working page by page, you can improve the project and its main indicators. Marketing efforts will become more effective, users will be more and more satisfied, and you’ll find more money coming in. As practice shows, it’s the improvements to the interface that often gives a good return on investment.
Technologies are imperceptible for the end user, but they’re a very important component for the project owner. Today, technologies become obsolete very quickly. New versions of languages and frameworks come out every two or three years, and CMSs are updated all the time. New versions are often very different from older ones, and offer significantly more possibilities. On the other hand, older versions are no longer supported, and it becomes problematic to work with them. It’s not a question about the capabilities of the new version, but rather about the loss of efficiency and development speed.
The technical platform of any Internet project needs to be constantly updated. To do this, follow a few important rules. First, develop everything correctly, taking into account future updates — in particular, don’t touch the CMS kernel, otherwise it won’t be updated. Secondly, write the code clearly and competently, taking into account existing quality standards (for example, one of the standards for PHP). Thirdly, it’s important to comment on the code so that the programmer, who updates the system, even if he didn’t create it, can understand it without any problems. Fourth, it’s important to create and maintain up-to-date project documentation, where the system will be fully described from a technical point of view.
Deploy both code integrity and a proper set of technologies. The larger the project, the more technologies can be used, but it’s important to choose technologies judiciously and not according to what the programmer knows, but according to what is really effective for resolving a particular problem. With an increase in the number of visitors, part of the project may need to be rewritten; everything cannot be scaled to fit, especially if the wrong architecture has been built in and there’s no possibility of scaling, which is quite common. Sometimes you need to rewrite the whole website if it was written immediately, say, on the CMS, and website traffic has grown and the site has begun to slow down or even crash. In any case, a good technical analysis is needed before making decisions on technologies.
Today, there are many different third-party services that are worth using. For example, the services of commodity recommendations for online stores, payment aggregators, login services through social networks, texting, various CRMs, etc. This means that the site owner often has a choice: to write his/her own unique system or use a third-party service. Both options can definitely be used. It’s important to understand that using a third-party service is cheap, but all the data is at this service, and we’re limited by its functionality. Therefore, such services are mostly used by smaller websites, while large ones develop their own unique systems. As a rule, projects are launched with a minimum number of services, but there are more and more of them as the website ages. For some reason or other, many people believe that it’s problematic to add such services. In fact, this is often a fairly simple task. Such services have their own API and good documentation, which makes them easy to add.
Servers and Infrastructure
It all depends on website traffic and how fast the project grows. If the site was created on an out-of-the box CMS, and its traffic starts to grow rapidly, then only by adding servers or moving the website to the cloud (such as AWS) will prevent the website from crashing. By the way, this is one of the most common problems when you economize on technologies and develop your website using a packaged solution, but after you must spend more on servers so that this option works in some way or another.
Everything is much simpler on large websites (on frameworks or in native language). You can perform load testing in advance and find out how many visitors the current servers will sustain. Two indicators are usually measured: standard and peak loads. Then, you calculate the loads, which will allow you to understand when, how and how many servers to add so that this option works without problems. People often start from one server, then a separate server for the database appears, then there are several more, etc.
If we don’t perform too many services or anticipate too many heavy loads, we’ll have almost nothing to do with the servers. You configure the website and automatic backup once, and occasionally check that everything works correctly. However, this is how small websites will operate, whereas professional server admins must work constantly with larger ones (they monitor security, loads, disk space and their operability so that they don’t crash, etc.).
You need to work constantly on content: improve old content and add a new one. This is especially true for information sites, stores, large portals, etc. Once again, many people make a basic mistake: instead of developing the content of the website, they fill it once and harvest customer feedback. This is wrong, and over time, such a website will receive fewer visitors, and, consequently, bring in less money.
Content strategy should be part of an overall strategy for developing the resource, which will indicate which content to use, how, where and when to publish.
We must remember that content is unique. Both users and search engines love unique content. In addition, the content must meet SEO requirements, so that it appears in the top results in the search engines and can stimulate more search engine traffic.
You should constantly motivate users towards further actions on websites where they generate the content in whole or in part; you must think it through and develop special tools. For example, if a person bought something in an online store, you can send him/her an email one or two weeks later and ask for feedback on the product.
Marketing and Promotion
At first glance, it seems that everything is very simple: Need more sales? Just buy more traffic, and it’s in the bag! In fact, we’ve got something to work on and improve here, like with the site itself. There’s a remarkable quote by John Wanamaker: “I know that half the money I spend on advertising is wasted. The trouble is that I don’t know which half.” It’s a pretty old idea — in the modern world there are many tools for measuring effectiveness. Moreover, we’re talking about Internet projects where everything can be measured and improved, and this should be used.
First, you need to prepare the website. The first thing to check is whether it’s been optimized for search engines. There’s a lot of information on SEO, for example, 20 rules for SEO-optimization of eCommerce projects. Next, check the content of the site: not only should it be optimized for search engines, but it should also motivate users to do something. It’s worth checking the interface separately: it should also prompt users to do something useful for the project. For contextual advertising, you need to create special landing pages to increase conversion, etc. There are a lot of things you can do when preparing the website, and almost all of them have an impact on conversion, the average purchase size, etc. In other words, you need to do this in order to get the most from users who visit the website.
When the website is ready, you can optimize the communication channels, through which potential customers enter the site. They can often be made more effective. To begin with, it’s necessary to calculate the economic efficiency of each channel, identify existing problems and possibilities for increasing efficiency, and then focus on optimizing and redistributing the budget from less effective to more effective ones. There’s much more work here than with the website, but this is a topic for another article.
Effectiveness Analysis and Strategy Adjustment
So, now that we’ve completed a tremendous amount of work on all fronts, our old website should begin a new life and become much more effective, often even at times, especially if no one had worked on it systematically before. However, this is still not enough. If you make such significant changes, you must constantly monitor the effectiveness of all corrective actions. It’s quite normal that not all assumptions will be effective. The team’s task is to find out exactly what has worked well and to what extent (in terms of profit), and make adjustments to the strategy on the basis of these results.
In Conclusion, What Should be Done with an Operational Website?
An Internet project is like a living organism. You need to work on it constantly and improve it. Once the project has been created, you’ll see if the project turns into a viable business or not. Moreover, if it’s clear that the idea’s a big hit, and the website has already started to bring in money or will generate more in the future, you need to start working on its development.
The type of project will determine where and how you start. Many websites are so outdated that they have to be completely redone. Large sites are regularly updated; small ones are substantially refreshed every three to five years. The most important thing is to move forward, and this process must be systematic and comprehensive. You can’t work on the website and forget about marketing, or vice versa. You can’t improve only one sphere, because the others will fall behind, and there will be no synergy effect. If you haven’t yet begun to actively enhance the effectiveness of your project, or the process has become very slow for some reason, then it’s time to tackle it head on.