Well-maintained product data is decisive for sales. The days, when customers relied on a single shopping channel, are over. Nowadays, they are reading up on products online and are researching carefully before they decide to buy. This is why it is so important to work with optimised product data. Especially in ecommerce, the information must therefore be meaningful, detailed and always up-to-date.
Information crucial to success
Regardless of whether it’s the possibility of even finding a product, information in a personally designed newsletter or an appealing product description – meaningful product and category information helps the customer to make decisions.
In product data marketing, vast amounts of product information are often processed and managed, used for campaigns and product text and transformed into personalised messages for customers. This also makes product data interesting for automation, because if you have a well-maintained data structure, you can also use it for automated text generation with the Text Tobot.
But not all product-related data that a company possesses is suitable for machine evaluation. In order for the Text Robot to actually use the data feed, it must meet certain criteria. Consequently, the product information must be prepared accordingly in order to obtain attractive and meaningful product descriptions by means of automated text generation.
Product data and what it can say about an item
Let us first look at the data that most companies have anyway. Often, master data from a PIM or other product data management tools is well suited for automated text creation. Master data is simple, descriptive information such as product names, manufacturers, dimensions etc.
A structured overview of this data (according to the model: product=x, manufacturer=y, length=a, width=b, height=c) already enables the creation of simple sentences, which may even be used for all products of a supplier. We can already create a short text like:
The smartphone from x is available in the size a x b x c.
The bottom line is that it would of course make more sense to specify a screen diagonal here, but you understand the principle. By the way, the screen diagonal – or any other surface – could also be calculated by the Text Robot from the dimensions, thus creating completely new possibilities. In the next step, we collect data on properties and features. Structuring these requires a little more effort and this is also the point where it often gets stuck.
Product features as structured data
Most companies have a wealth of data on product features and properties. This is important, because who buys a smartphone with the above product description?
But how can the data be structured to be machine-readable? Since the assignability of data plays an important role for the Text Robot, it is important to split the product data into unique data fields. This could be, for example:
This information can then be used to create a few more meaningful sentences:
The Android smartphone with a touch screen has an 8MP camera. With LTE data transmission, you will be quickly connected when you are on the move.
The individual data fields may apply to several products and may contain data of completely different products for another data set. For example, the data field “operating_system” for a desktop computer could contain the information Windows 10.
Meta data for an interpretation of information
Of particular interest to large ecommerce retailers is also the so-called meta-data, which enables the Text Tobot to work interpretatively. Such meta-data can be, for example, sales figures, return rates or evaluations of other buyers.
Accordingly, a low return rate indicates, for example, that customers are very satisfied with the product, which enables the Text Robot to write a sentence like:
A particularly popular product that is guaranteed to provide you with joy for a long time.
Other interesting sources of information for automated text creation can also be shop information and manufacturer data, which offer further added value for automatically generated text.
However, it is and remains important to structure the data optimally so that it is machine-readable. Data fields that contain several badly separated attributes are often unusable and also fields with continuous text cannot be evaluated by the Text Robot.
Product data optimisation for automated content creation
But what does the optimal and purchase-relevant data for the automated creation of product descriptions look like? If I want to prepare and optimise my product data for content automation, there are a few things to consider. Specifically, this means that I have to take a close look at my data and separate it if necessary.
In many cases, online shop owners have, for example, a product description or a product name that already contains the most important features. In the name of a sofa, for example, the colour, material and perhaps even the dimensions are already included. Even though this is practical for the header and the customer thus can see immediately all important information about the product, it is useless for the Text Robot.
Basically, such a data field contains a lot of wasted information – simply because the attributes from this data field are not machine-readable. But if I have a separate data field for each attribute (“colour”, “material” or “dimensions”), I can generate a separate sentence for each attribute. If this is then displayed in different variants and different sequences, variance arises and duplicate content is avoided.
The keyword here is granularity and the more granular the data, the more flexible and multi-faceted the text can be.
Frequency of occurrence or level of completeness
Of course, there is no point in creating hundreds of data fields, which then only affect one product in thousands. For automated content creation, it is a matter of weighing up the pros and cons: Which data fields describe most of my products and what is the level of completeness when I create specific attributes?
This train of thought is not one-dimensional, because in some cases it can make sense to use an attribute that is only slightly filled in/used – namely if exactly this attribute is a special feature that I want to emphasise in my automatically generated text.
The application with shimmering pearls makes this T-shirt a real eye-catcher.
Even if there are very few T-shirts with such an application in the shop, this category (T-shirts) is especially important to me.
Granularity and structure
To begin with, create 20-25 attributes and increase their number to the rule-of-thumb value of about 90 attributes for the complete text project. With this amount, a significant variance can already be created. The more attributes that are filled in correctly I have, the more flexible and detailed will the automated text I can generate be.
It may be a bit time-consuming to prepare the existing product data for automated text creation with a Text Robot, but it is worth it! Because with well-structured and optimised data, great product descriptions can be created that appeal to and inspire customers.
Video: Types and use of data
The CEO of AX Semantics, Saim Alkan, has summarised the topic of product data very nicely in a video.
Large companies and corporations usually manage the digital transformation on their own. They put together digitalisation teams headed by a Chief Digital Officer, set up start-ups and recruit digitally literate staff. Above all: financial resources are usually the least of all bottlenecks – they are hardly dependent on subsidies or funding programmes.
Small and medium-sized enterprises (SMEs) and handicraft companies have a much harder time. Building up own resources and expertise is usually not economically feasible. In addition, their daily business forces them to focus only little or not at all on issues related to digitalisation.
The result speaks volumes: SMEs want to digitalise. We can help.
Politics and government have recognised this imbalance. The ‘go-digital’ subsidiary programme was launched under the auspices of the Bundesministerium für Wirtschaft und Energie (‘Federal Ministry of Economics and Energy’ of Germany), BMWi for short. The subsidiaries provided are available to medium-sized and small companies to enable them to tackle and implement digitalisation projects more easily and efficiently.
What can be subsidised with ‘go-digital’?
Roughly speaking, SMEs and handicraft companies in cooperation with competent consultancy firms authorised by the Federal Ministry of Economics and Energy can take advantage of subsidiaries and implement projects in the following core areas:
1. Support in IT security
Risk and security analysis to evaluate threats and possible weaknesses within the company’s IT. This also applies to new purchases of IT and telecommunications infrastructure.
Initiatives and measures that serve to improve the security management of internal IT.
The aim of this module of the programme is to better protect small and medium-sized enterprises against cybercrime, which is a potential threat during digitalisation.
2. Support in digitalisation of business processes
Introduction of software solutions and tools that improve, simplify and accelerate business processes between companies, business partners and customers.
Digitalisation of internal core processes of departments such as marketing, sales, logistics, production etc.
The aim of this module is to digitalise work processes as completely as possible and improve online processes.
3. Advancements in the field of digital market development
Establishment of a strategy for online marketing and sales.
Programming of a legally compliant and modern website/representation, ideally in connection with an online shop. This also includes all downstream services such as content production, social media or website monitoring.
Setting up digital processes for the management of the web shop, electronic payment transactions or dispatch of goods.
Within this module, the subsidiary programme aims to support SMEs and handicraft companies in setting up online marketing and sales of goods and services via the Internet.
Your company can benefit from the subsidy rate of 50% per daily rate of € 1,100. The grant can be taken up to 30 days within six months.Expressed in concrete figures, for you this means: A grant of up to € 16,500 for your digitalisation projects.
Secure subsidiaries now!
Who can take advantage of the ‘go-digital’ funding programme?
The conditions for benefiting from this programme can be listed quickly. Support is given to legally independent small and medium-sized enterprises in the commercial or handicraft sectors. The following applies:
Your company must have fewer than 100 employees when entering into the contract. Includes partner enterprises and associated enterprises.
The turnover or balance sheet total of your company in the previous year does not exceed € 20 million.
The permanent establishment or branch of the company is located in Germany.
If the company is a “partner company” or “affiliated company”, the consulting company applying for funding must confirm the eligibility under the de minimis regulation of the European Union with regard to the number of employees and annual turnover/balance sheet total.
How do we support you with ‘go-digital’?
uNaice is a consulting company certified by the Federal Ministry of Economics and Energy. We are thus authorised to apply for subsidies together with you.
Which digitalisation project suits you best?
We can digitalise your content production and automatically generate unique and SEO-relevant text from data.
Online shop owners often wonder how they can make the shopping experience as pleasant as possible for potential customers. Great importance is attached to the visual presentation of product descriptions and product pages. The use of large format images or videos is becoming increasingly important.
In addition, the visitors of the online shop are constantly measured in the background. Their click paths and mouse movements are evaluated. Based on the analyses, the UI (user interface) and on-page elements are permanently optimised.
All these efforts have one goal: to increase the rate of buying customers. Because the higher the conversion rate, the longer the cash tills of the online shop keep ringing. However, in addition to the purely technical view, an essential element of the product pages is often neglected: the product description, i.e. the product text itself describing the goods.
Search engines love good product descriptions
The product description, i.e. good product content, is one of the most important factors for increasing the conversion rate – and thus sales. There are several reasons for this:
Good product descriptions inform customers about the characteristics, features and added value of the product. Product descriptions with added value can be the tip of the scale. Especially if the customer is still uncertain about the purchase.
Product descriptions written with the potential target groups in mind, create the necessary trust and provide interested parties with the product information that is relevant to them.
Thus, well-prepared product descriptions demonstrably contribute to a better ranking in the search results on Google. As a result, they also increase the conversion rate.
But a glance at some online shops shows all too often how carelessly the description of the products is handled. Because many online retailers are content with enumerating product features in the form of boring bullet points or cryptic continuous text instead of creating good text that describe the product in detail.
Even completely outdated data or even missing information are unfortunately a daily occurrence. Important questions of the readers are left unanswered. Added values such as tips, benefits or features of the product are also often missing. It could be even worse: Even overview pages often lead a shadowy existence or are filled with crude text. This discourages customers and, in doubt, they prefer not to buy or look around at the competition.
Product descriptions of the manufacturers are not a good alternative
Even worse: Many online shop owners, in their precarious situation, default to the product descriptions provided free of charge by the manufacturer. However, this has disastrous consequences.
Integrating suppliers’ product descriptions in web shops is a risky undertaking.
Let’s put it to the test. Let’s take an extract from any product description of a manufacturer and start a Google search. The result is frightening: there are over 3,200(!) hits. These include a considerable number of online shops. As a result, any individuality and uniqueness falls by the wayside. For this reason, many online shops disappear in a mass of identical search results. We can deduce three things from the example:
The use of manufacturers’ product descriptions in online shops robs retailers of any chance of positioning themselves. Moreover, the individual approach to the target group suffers as a result.
In short: Those who increasingly use manufacturers’ product descriptions expose themselves to a risk that should not be underestimated. the Google search algorithm does not like the multiple use of so-called duplicate content at all. Thus, a generally worse ranking in the search results is the consequence.
Using manufacturers’ product descriptions weakens the conversion rate instead of increasing it.
Conversion-strong product descriptions not without effort
There are numerous reasons why product descriptions are neglected. For online retailers, the creation of product descriptions that are appealing to the target group, unique and thus SEO-relevant, poses a number of challenges:
Speed: A lot of time goes by until the descriptions have been created. This means that product descriptions can often only be added to the product when its half-life has long since passed.
Quality: The production quality of externally commissioned text is – to say the least – often horrible. Furthermore, the requirements of conversion-strong and SEO-relevant text have not been fulfilled at all. On top of that, all too often you purchase product descriptions with orthographic and grammatical inadequacies. It takes a lot of extra time to eradicate the errors.
Effort: Producing your own high quality product descriptions is a costly and maintenance-intensive matter. In low-margin online business, however, the retailer must think cost-effectively.
Sustainability: The constant maintenance of category pages is often ignored and not carried out at all. The effort required by the constantly changing product range is simply too high. Product pages and overview pages are often completely forgotten.
Good product descriptions demonstrate their effectiveness
But no matter how you twist and turn it, one thing remains undisputed: good product descriptions are effective and boost sales. To achieve this, the target group must be addressed accurately. The adequate tonality of the texts helps in this case. With SEO-relevant product descriptions the cash tills will keep ringing significantly, because those willing to buy will find the right product. The reason for this is an increased conversion rate. The next post will explain how the challenges described above can be met efficiently, sustainably and economically by using a Text Robot.
Further contributions on the topic of Text Robots:
Our Text Robot is a solution that automatically writes a large number of copies at low cost and in a short time, on the basis of existing data. In our discussions with customers, we repeatedly emphasise that a Text Robot is not an artificial intelligence, but a machine that is programmed and trained by humans. Comparable to a welding robot in automobile production. The machine does not learn automatically, but completes human actions faster, more precisely and more efficiently, without becoming tired or making careless mistakes.
This leads to the question: At what point does a company need a Text Robot?
Let’s stay with our example of the industrial robot: If you have to apply welding seams five times a week on five different cars in ten different places – is it worthwhile for you to purchase a welding robot for a high six-figure sum? Probably not.
If you want to set the same welding spots 500 times a day on the same car in the same high quality, the purchase is of course already worthwhile.
The situation is very similar with our solution for automatic text creation.
If you need hundreds of copies per month or several times a year in consistent quality, based on structured data, then a Text Robot is a very good – and above all cost-effective – solution for you. Examples are descriptions created from product data, individualised customer letters from customer data or personalised newsletters from CRM data.
Of particular interest is the addition of third party data, which generates an additional service or knowledge gain for your customers when integrated into your company’s text. This can be weather data, for example, but also location-based information, industry data, survey results and so on.
How does uNaice proceed with the configuration of a Text Robot?
The first step is an analysis of your initial data. We can then make reliable statements as to how many of your data records we can at least provide with a text. The minimum text consists of two main sentences and two subordinate clauses covering as many data categories as possible.
This is followed by an investigation into which data categories should be described in depth with longer text. Criteria that are directly related to your business model will also be used here. For example, which data groups have the greatest potential for turnover, conversion increase or search engine optimisation?
Once these decisions have been made, our computer linguists and developers get to work equipping and programming the machine individually with text modules, synonyms, regional and industry specifics and much more. At this point, native speakers are also used to lay the foundations for each of the currently 119 possible languages that the Text Robot will be able to use later. The result is a machine that, at the push of a button, will generate text for you that sounds as if it has been written by a human, and which will then be imported into your systems via a REST-API.
When does the generation of text with a robot make sense?
The simple answer is:
“As soon as automatically generated text in large numbers and within a short time would be a competitive advantage for your company”.
Basically, from our experience, we can say that so far we have been able to uncover competitive advantages with every company with which we have started looking into existing data. Often with a direct impact on day-to-day business. After a short training period in the customer business, we often also found worthwhile new business opportunities that nobody had thought of yet. Rapid internationalisation (in up to 119 languages in which the Text Robot can be trained) allows new markets to be developed.
Further contributions on the topic of Text Robots:
SEO experts worldwide have long debated whether the release of thousands upon thousands of new or additional websites could have a detrimental effect on Google’s ranking. For example, what would happen if a company launched 50,000 new or additional product descriptions? This question is not unreasonable. Customers who rely on automated text creation and use a Text Robot for this are often faced with this consideration.
There has been much debate in professional circles about the possible effects. Is it good to put so much new content online? Does it have a rather negative effect on the ranking? Or does it not matter at all?
Since there has never been an official statement on the part of Google, the credo was: better to play it safe and publish in several steps – better safe than sorry.
A clear statement for the first time
Now, John Müller, Google manager and one of the search engine giant’s top webmasters, has taken a stand on precisely these cases in an expert question and answer session. When asked whether it was OK to publish 100,000 web pages at once, he said: “Sure, no problem!”
From an SEO point of view it is absolutely fine to publish everything at once. The only thing that needs to be ensured is that the new content is detectable by Google and that the web server can handle the queries of the Google bot.
“In one go” should be a clear preference
Furthermore: Müller is against a step-by-step publication. The structure of the search engine index would be less problematic overall if it did not have to be put together piece by piece by new results from the bot.
One thing is thus clear: If you rely on automated text creation and have a Text Robot produce a lot of fresh and unique content, you can publish it immediately. And that is exactly how it is intended.
You can find the interview with the question and Müller’s answer in this video (from 19:12).
I think the the biggest impact, the biggest effect you’ll see is that Googlebot will try to crawl a lot of pages. Especially if it’s a larger well-known site, where we already know that crawling pages and indexing pages from that site makes a lot of sense. Then we’ll probably kind of try to crawl a lot of pages on your server and it will be good if your server had the capacity to deal with that.
But from an SEO point of view that’s generally not an issue.
Some sites put their archive online all at once. I think kind of artificially introducing a kind of a trickle into the index is something that often causes more problems than it solves anything. So obviously putting a hundred thousand pages online all at once doesn’t mean that these pages will rank all at once immediately. It’s still going to take time for us to actually rank those pages properly in the search results. But I don’t see anything against putting them all on online.
Further contributions on the topic of Text Robots:
At the latest when setting up a Text Robot, customers regularly ask themselves whether and in what way the success of using a Text Robot can be measured. There are at least four answers to this question:
1. Amount of text produced in relation to time
It is helpful to regularly generate a visualisation during the course of the project that shows how many copies have already been automatically generated by the Text Robot within a period of time. A so-called burn-down diagram, optionally also a burn-up diagram, is quick and easy for everyone to grasp. It shows in the x-axis the applied time and in the y-axis the number of automatically created copies.
To underline the success of speed in text production, this can be put in relation to the manual production of text. A copywriter manages between 5 and 10 product descriptions per day. With the help of such diagrams the extremely high scaling effect of a Text Robot becomes visible.
2. Money saved with automated text creation
With the help of our amortisation and scaling calculator, after entering the optional figures, you can see how quickly the investment will pay for itself and what economies of scale will be achieved.
Usually a product description created by a text agency costs around € 0,13 per word or between € 7 and € 22 per product text. We have also met customers who pay € 25 ore even more per text. This is initially only the pure purchase cost of a text.
In the case of a full-cost accounting, there are a lot of additional items: time and effort for briefings, receipt of the copies, editing, coordination and approval loops as well as the setting of the text, for example, are very important as well. Including these surcharges, costs of € 17 to € 45 are easily incurred – depending on the characteristics, requirements and structure of the text.
It is advisable to convert the automatically generated amount of text into costs that have already been “saved”, thus making it clear how much money no longer needs to be transferred to text agencies or no longer accrues in your own company. In the event that this money has not yet been spent because no text has been written to date, these figures can be used to show the costs saved compared to manual text creation.
We have programmed an amortisation calculator which – fed with the key figures – calculates from which month on the investment into a Text Robot has fully paid off. Please get in touch with us for an individual calculation.
3. Development of the conversion rate
Text that optimises itself on the basis of reading behaviour and thus becomes increasingly attractive is no longer a science fiction scenario.
Ultimately, all the data that can be collected with Google Analytics, is possible. Within a separate (and shared with the customer) Google Analytics account, evaluations can be generated that show a correlation between the displayed text and the target, such as the purchase (shopping basket). Text that optimises itself on the basis of reading behaviour and becomes more and more attractive with the help of the Text Robot is no longer a science fiction scenario.
Two text variants are generated by the Text Robot for each product and are displayed in an A/B test procedure. If, for example, text variant A prevails, you naturally want to produce more text based on the better variant.
The good news: In a further expansion stage, the knowledge gained from the A/B test is fed back to the Text Robot and new text variants are generated under the parameters recorded. This creates “a rule cycle” that automatically leads to better and better text over time. Self-optimising text – controlled by the reading behaviour of the users – will thus be created. This is an absolute novelty.
4. SEO and visibility in search engines
This part is very complex because both web visibility and search engine ranking depend on a myriad of factors and not even a good SEO consultant knows the algorithm used by Google down to the last detail.
Furthermore: popular measuring instruments like Sistrix (shows the visibility index in search engines) only draw on a small subset of keywords. The high variance of the automatically created text is not necessarily reflected here.
However, it can be expected that in cases where the previous product text was short or with a lot of “duplicate content” (marketing text of the manufacturers that are used as a product description), the visibility and ranking will benefit. It is advisable for an SEO specialist to accompany the changeover to automated text creation and – where necessary – to set the right course.
It is absolutely necessary to observe the following: Only a clean zero measurement and an analysis procedure documented before the project can lead to true future statements and comparative values. Changes during or after the project lead to misinterpretations and false comparative values.
Further contributions on the topic of Text Robots: