Product Descriptions


  1. Pro­duct descrip­ti­ons for eCom­mer­ce pro­vi­ders
  2. Why pro­duct descrip­ti­ons are important for Search Engi­ne Opti­mi­sa­ti­on (SEO)
  3. What makes a good pro­duct descrip­ti­on?
  4. Uni­que Con­tent vs. Dupli­ca­te Con­tent
  5. How to Effi­ci­ent­ly Crea­te Pro­duct Descrip­ti­ons
  6. How to Auto­ma­ti­cal­ly Gene­ra­te Pro­duct Descrip­ti­ons
  7. Ama­zon-SEO with no addi­tio­nal effort
  8. Trans­la­ting pro­duct descrip­ti­ons bet­ween lan­guages



Product descriptions for eCommerce providers


Pro­duct descrip­ti­ons are among the most important texts in eCom­mer­ce. The qua­li­ty of a pro­duct descrip­ti­on con­tri­bu­tes as much to sales suc­cess as pri­ces, pic­tures, or posi­ti­ve custo­mer feed­back. 

A good pro­duct descrip­ti­on increa­ses the likeli­hood of a purcha­se being made. But they have many other advan­ta­ges for online retailers.


Why product descriptions are important for Search Engine Optimisation (SEO)?


All online retailers know the impor­t­an­ce of pla­ce­ment in Goog­le search ran­kings. The more opti­mal the inter­ac­tion bet­ween tech­ni­cal and con­tent fac­tors of a page, the hig­her the search engi­ne pla­ces an ent­ry in the result lists.

Mobi­le usa­bi­li­ty, loa­ding time or inter­nal lin­king are important cri­te­ria for the pla­ce­ment of a page in the search engine—structure and qua­li­ty of con­tent are also decisi­ve fac­tors. Texts such as pro­duct descrip­ti­ons are so cen­tral to Goog­le becau­se the algo­rithms that com­pi­le ran­kings can read and eva­lua­te text bet­ter than other infor­ma­ti­on, such as images.

Key­word-opti­mi­sed, well-writ­ten texts have a bet­ter chan­ce of appearing at the top of an inter­net search. Goog­le also takes into account the length of a text and users appre­cia­te com­pre­hen­si­ve­ly-writ­ten pro­duct descrip­ti­ons.

The bet­ter and lon­ger a pro­duct descrip­ti­on is, the lon­ger a user stays on the site. The­re are two aspects to this: With increa­sing length of stay, the pro­ba­bi­li­ty that an inter­ac­tion will take place increa­ses, such as a purcha­se or a sub­scrip­ti­on.

The amount of time that peop­le spend on a web­site is also important for SEO. The lon­ger a con­su­mer stays on the site, the stron­ger the signal to Goog­le that the con­tent on that site is rele­vant. This reads to a long-term reward for the site against poor­ly-per­for­ming com­pe­ti­tors.


What makes a good product description?


The­re is no model for good pro­duct descrip­ti­ons. Ins­te­ad of fol­lo­wing a tem­pla­te, it is advi­s­able to con­si­der cer­tain aspects. Depen­ding on the pro­duct, one fac­tor or ano­t­her may car­ry more weight.

  • Text Length: A con­ti­nuous text should be around 200 wor­ds, alt­hough this can be lon­ger. The more explana­ti­on is nee­ded, the lon­ger the text should be. Each sen­tence should be no lon­ger than 25 wor­ds.
  • Wor­ds and Phra­ses: The wor­ds used in the pro­duct descrip­ti­on depend on the tar­get audi­ence. Facts are par­ti­cu­lar­ly important. If a cos­metic arti­cle or pie­ce of fur­ni­tu­re is being pre­sen­ted, a rich descrip­ti­on that prompts moods and emo­ti­ons is hel­pful.
  • Custo­mer Bene­fits: How does the pro­duct help con­su­mers sol­ve their pro­blems? How does it save time and/or money? How does it make them hap­pier? A good pro­duct descrip­ti­on must ans­wer the­se ques­ti­ons.
  • Com­pa­ri­sons: This is par­ti­cu­lar­ly important if simi­lar models are avail­ab­le as it helps users deci­de.
  • For­mat­ting: The lon­ger the text, the more important it is to have para­graphs to break it up visual­ly. Bold or ita­li­ci­zed terms and phra­ses help the user to grasp the pro­duct text more easi­ly. Tip: For­mat­ting com­bi­na­ti­on key­wor­ds is a pro­ven trick for activating SEO poten­ti­al for niche terms with low search volu­me.


Unique Content vs. Duplicate Content


Con­ti­nuous­ly crea­ting pro­duct descrip­ti­ons is a head­a­che for online retailers. When a new collec­tion comes into an online fashion shop, nume­rous pro­duct texts for jackets, trou­sers or pull­overs have to be crea­ted and ent­e­red wit­hin a very short time. Updating exis­ting pro­ducts, such as the annu­al ver­si­ons of soft­ware, also means that a pro­duct descrip­ti­on must be adap­ted.

In many indus­tries, the manu­fac­tu­rer pro­vi­des a descrip­ti­on text in addi­ti­on to pro­duct images and struc­tu­red data such as RRP, size or colour. The shop ope­ra­tor can make it easy for him­s­elf by adopting the manufacturer’s pro­duct descrip­ti­ons word-for-word.

The­se ready-made texts and struc­tu­red data can be auto­ma­ti­cal­ly impor­ted into shop sys­tems such as Magen­to, Shop­ware or xt-Com­mer­ce. The advan­ta­ge is that the­se descrip­ti­ons go ‘live’ imme­dia­te­ly without much invest­ment in time or money. The dis­ad­van­ta­ge is that the­se descrip­ti­ons are not the ‘uni­que con­tent’ that Goog­le gra­vi­ta­tes towards, which signi­fi­cant­ly redu­ces the chan­ces of SEO.

In prac­tice, the manufacturer’s text appears on the web­sites of many dif­fe­rent online mer­chants and mar­ket­pla­ces, inclu­ding Ama­zon. Ins­te­ad of being ‘uni­que con­tent’, this is ‘dupli­ca­te con­tent’. Not having uni­que con­tent does not necessa­ri­ly have to harm a retailer, but the use of pro­duct descrip­ti­ons from the manu­fac­tu­rers will not have a posi­ti­ve effect under any cir­cum­s­tan­ces.


How to Efficiently Create Product Descriptions


It is not com­mon for online retailers to employ peop­le to pro­du­ce con­tent, gene­ral­ly far­ming this out to fre­e­lan­ce aut­hors or agen­ci­es. This work is char­ged using a varie­ty of payment models, such as packa­ge pri­ces for lar­ge num­bers of texts or bil­ling per word.

Lea­ving the qua­li­ty of the text to one side, con­ven­tio­nal copy­wri­ting often invol­ves addi­tio­nal work apart from the fees. Coor­di­na­ti­on, brie­fing, proofrea­ding, and inte­gra­ting the descrip­ti­ons is usual­ly car­ri­ed out by an employee.

Some of the­se steps can be eli­mi­na­ted by using text gene­ra­ti­on. The case for using a SaaS solu­ti­on is intui­ti­ve. The basic form of such a solu­ti­on is for gaps wit­hin tem­pla­tes to be fil­led with data. This data could encom­pass such things as brand, colour, size, con­di­ti­on, or pri­ce, and come from struc­tu­red data.

This struc­tu­red data is a pre­re­qui­si­te for auto­ma­tic text gene­ra­ti­on. As a rule, this should come from the mer­chan­di­se manage­ment sys­tem or from manu­fac­tu­rers. It is ren­de­red as a CSV or XML file into the online sys­tem. From the­re, it can be uploa­ded, manu­al­ly or through an API, to a gene­ra­tor. The sys­tem then auto­ma­ti­cal­ly crea­tes con­tent using intel­li­gent lin­gu­is­tic ana­ly­sis.


How to Automatically Generate Product Descriptions


The struc­tu­re of the respec­tive pro­duct descrip­ti­on is defi­ned by tem­pla­tes and con­di­ti­ons. Tem­pla­tes are essen­ti­al­ly varied gap texts with a multi­tu­de of syn­onyms, adverbs and other lexi­con ent­ries. Con­di­ti­ons are the cir­cum­s­tan­ces that must be ful­fil­led for the tem­pla­te to be used.

The pro­cess is done via a soft­ware engi­ne that arran­ges the tem­pla­tes in a cer­tain order. The given order in which the tem­pla­tes are put toge­ther is cal­led a ‘sto­ry plot’ or ‘nar­ra­ti­ve’. Tem­pla­tes and con­di­ti­ons must be defi­ned by human edi­tors wit­hin the frame­work of an initi­al set­up; the solu­ti­on then works inde­pendent­ly.

Depen­ding on the ori­en­ta­ti­on of the shop, the copy­wri­ters can crea­te the pro­duct descrip­ti­ons in dif­fe­rent tones. It is pos­si­ble to pro­vi­de casu­al, emo­tio­nal, or tech­ni­cal-soun­ding messa­ges to custo­mers.


Amazon-SEO with no additional effort


Ama­zon, eBay, and other online retailers are a dou­ble-edged sword for online mer­chants: high reach and signi­fi­cant sales on the one hand, but com­pe­ti­ti­ve pres­su­re and pri­ce war on the other. Only tho­se who achie­ve a high level of visi­bi­li­ty on plat­forms and mar­ket­pla­ces, per­fect­ly pre­sen­ting their own offers in text form, can sur­vi­ve in the­se chan­nels in the long term.

Auto­ma­tic text gene­ra­ti­on via a SaaS solu­ti­on allows the pro­duc­tion of opti­mi­zed, uni­que texts per chan­nel in a mat­ter of seconds. At the push of a but­ton, soft­ware crea­tes mul­ti­ple varia­ti­ons of a source text. Via an inter­face, one of the­se ver­si­ons is impor­ted into the sys­tem of the respec­tive plat­form.

Spe­cial pro­mo­ti­ons and cam­pai­gns, espe­ci­al­ly on the occa­si­on of the high-tur­no­ver ‘Black Fri­day’ and ‘Cyber Week’ on Ama­zon, can also be tar­ge­ted more spe­ci­fi­cal­ly with the help of natu­ral lan­guage gene­ra­ti­on.


Translating product descriptions between languages


Online retailers have more issu­es out­side of con­trol­ling their pro­ducts wit­hin their own sales are­as. Many retailers work across bor­ders and con­se­quent­ly need to pro­du­ce pro­duct descrip­ti­ons in mul­ti­ple lan­guages.

A mul­ti­lin­gu­al SaaS text gene­ra­ti­on solu­ti­on is a power­ful tool for a con­sis­tent brand pre­sence. Auto­ma­ted trans­la­ti­on bet­ween lan­guages is important across Euro­pe, mea­ning that sys­tems should be able to pro­du­ce in English, French, Ita­li­an, Ger­man, Dutch, and Spa­nish. Bet­ween cer­tain lan­guage pairs, such and English and Ger­man, machi­ne trans­la­ti­on is now so accu­ra­te as to pass mus­ter with an human edi­tor.

If a gene­ra­tor is pro­du­cing text and lan­guage in mul­ti­ple lan­guages, the­re is a signi­fi­cant savings poten­ti­al for inter­na­tio­nal com­pa­nies. Trans­la­ting inter­faces such as Goog­le Trans­la­te and DeeplL are imple­men­ted in offe­rings such as