The Shop Owner With Excel and Google Analytics

interviews · Amfiteatru Economic · la publicación, 1 may 2026 · gratis

Lo que 25 dueños de tiendas en línea dijeron de la analítica y la IA: menos intuición y más ventas, según su propia percepción

El estudio entrevistó a 25 gerentes y dueños de pymes de comercio electrónico en Rumania. No midió resultados financieros reales ni probó que las herramientas causaran las mejoras.

Versión breve · la versión detallada sigue, unos 7 min

Pregunte al chat de weeklyAI

Pregúnteme por este estudio: a quiénes se estudió, qué encontró y qué no dice.

Las conversaciones se guardan mientras exista weeklyAI, para mejorar la publicación. Se responde en el idioma en que usted escribe.

El estudio, de un vistazo
Quiénes: gerentes y dueños de pymes de comercio electrónico
Cuántos: 25 entrevistas
Dónde: Rumania
Cuándo: septiembre y octubre de 2025
Tipo de estudio: entrevistas
Quién lo hizo: Universidad de Craiova
El límite que importa: No midió resultados financieros ni comparó con quienes no adoptaron; son percepciones.

Cómo se repartieron las 25 entrevistas por tamaño de empresa

Microempresas52%

Pequeñas empresas28%

Medianas empresas20%

Porcentaje de las 25 entrevistas realizadas a pymes de comercio electrónico en Rumania, según el tamaño de la empresa; no representa a todas las pymes del sector.

Herramientas que usaba cada tamaño de empresa

Microempresasfrente aPequeñas empresas

Las micro usaban sobre todo herramientas gratuitas o de bajo costo; las pequeñas sumaban soluciones por suscripción como CRM y correo marketing.

Pequeñas empresasfrente aMedianas empresas

Las medianas mostraban la adopción más avanzada, con sistemas integrales de gestión, plataformas de inteligencia de negocio y modelos predictivos de IA.

Lo que este estudio deja en sus manos es un espejo, no una promesa.
Lectura de weeklyAI

Ver · 0:54

"Usamos Google Analytics y los reportes de la plataforma. Armamos los datos en Excel". Así describió su operación el dueño de una microempresa, uno de los 25 gerentes y propietarios de pymes de comercio electrónico de Rumania que dieron entrevistas por teléfono entre septiembre y octubre de 2025.

El estudio, de investigadores de la Universidad de Craiova, encontró un patrón según el tamaño de la empresa. Las microempresas se apoyan en herramientas gratuitas o de bajo costo: Google Analytics, los reportes de la plataforma, Excel. Las pequeñas suman CRM y Power BI. Las medianas llegan a sistemas ERP y modelos predictivos de inteligencia artificial.

Los entrevistados dijeron que después de adoptar estas herramientas tomaban decisiones más rápidas y con más confianza. También reportaron ventas más estables, mejores márgenes por categoría, menos gastos en promociones y un manejo más eficiente del inventario. Lo que más mencionaron fue el cambio en la organización del trabajo: menos tareas repetitivas, más personalización. Un gerente de una empresa mediana lo resumió: "Las tecnologías inteligentes optimizaron los procesos de análisis y planificación".

Pero muchos de los participantes, sin importar el tamaño de su empresa, dijo que sus empleados carecían de las habilidades necesarias. Las microempresas señalaron el presupuesto limitado como el obstáculo principal: "El presupuesto limitado no permite soluciones avanzadas". Las empresas más grandes enfrentaron sobre todo la dificultad de integrar los sistemas y el tiempo que exige probarlos y ajustarlos.

El estudio no midió los resultados financieros. Registró lo que los dueños percibieron. Tampoco comparó empresas que adoptaron estas herramientas con empresas que no lo hicieron, así que no puede decir que las herramientas causaran las mejoras. Y todos los entrevistados eran pymes de comercio electrónico en Rumania: los costos, las reglas y los proveedores en América Latina, Estados Unidos o Canadá pueden ser distintos.

El estudio menciona que los costos altos y la falta de dinero frenan a las empresas, sobre todo a las microempresas. Los investigadores eligieron a las empresas a propósito, buscando pymes de comercio electrónico de distintos tamaños y regiones de Rumania.

Antes de pagar por herramientas que prometen más ventas, revise dos cosas en su propio negocio: si alguien en su equipo puede usar los datos que ya tiene, y cuánto puede gastar sin quedarse sin margen.

Qué significa para usted

Lo que estos dueños describieron no prueba que las herramientas causaran las mejoras, y todavía no hay razón para pagar por promesas de más ventas. Mientras tanto, fíjese en algo concreto: si alguien de su equipo ya sabe leer los datos que su negocio genera todos los días, y cuánto puede gastar sin apretar el margen. Esos dos puntos le dirán si está listo para dar el siguiente paso.

Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650

Quién pagó: El artículo dice que la investigación se realizó con el apoyo del Centro de Investigación Interdisciplinaria en Ciencias Económicas y Sociales, INCESA (Infraestructura de Investigación en Ciencias Aplicadas), Universidad de Craiova; no dice si los financiadores tuvieron alguna influencia ni quién prestó equipos o software.

Versión detalladaLos pasajes copiados del artículo, las ilustraciones y cada fuente con cuánto leímos de ella · 1343 palabras · unos 7 minLeerla →Cerrar

Lo que 25 dueños de tiendas en línea en Rumanía dijeron sobre analítica e inteligencia artificial

Un estudio cualitativo no mide ganancias: recoge cómo vivieron los gerentes la adopción de estas herramientas, y qué los frenó.

Así podría verse · ilustración generada por weeklyAI.watch, no es una fotografía

Veinticinco gerentes y dueños de pequeñas y medianas empresas de comercio electrónico en Rumanía fueron entrevistados entre septiembre y octubre de 2025. Los investigadores preguntaron qué herramientas usaban, por qué las adoptaron, qué obstáculos encontraron, cómo cambió su forma de decidir y qué efectos notaron en el negocio. No hubo grupo de comparación ni medición independiente de resultados financieros: lo que se recogió fueron percepciones.

Los entrevistados describieron una madurez digital intermedia, dominada por el análisis descriptivo —mirar lo que ya pasó—, mientras que las aplicaciones de inteligencia artificial estaban solo parcialmente implementadas¹. Los dos frenos principales que mencionaron fueron la falta de competencias digitales y las restricciones financieras, con diferencias según el tamaño de la empresa¹. Cuando esos obstáculos se superan, según los propios gerentes, la adopción contribuyó a fundamentar mejor las decisiones y a mejorar el desempeño financiero mediante optimización de ventas, reducción de costos y eficiencia en el manejo de inventario².

Vale la pena entender qué son estas herramientas antes de sacar conclusiones. La analítica de datos es el uso sistemático de técnicas estadísticas y analíticas para generar información útil al proceso gerencial; la inteligencia artificial, en cambio, designa tecnologías capaces de simular procesos cognitivos, como el aprendizaje automático y la generación de predicciones³. En la práctica, eso va desde mirar cuántas visitas recibió la tienda hasta pedirle a un modelo que estime cuánto de un producto se venderá el mes próximo.

Las 25 empresas no se parecían entre sí, y esa diferencia ordena casi todo lo demás. Las microempresas se apoyaban sobre todo en herramientas gratuitas o de bajo costo: Google Analytics, los reportes estándar de su plataforma de venta y Excel para análisis financieros básicos⁴. Las pequeñas sumaban soluciones por suscripción, como sistemas de gestión de clientes y plataformas de correo marketing⁵. Las medianas mostraban la adopción más avanzada, con sistemas integrales de gestión, plataformas de inteligencia de negocio y modelos predictivos para estimar demanda, ajustar precios y personalizar marketing⁶.

Los obstáculos también se repartieron de forma distinta. Los costos altos del software profesional, la complejidad de integrar sistemas entre sí, la necesidad de pruebas y ajustes repetidos y la falta de experiencia interna aparecieron como los principales problemas⁷. La diferencia de fondo: las microempresas tropiezan primero con el dinero, mientras que las empresas mayores chocan con la implementación y la optimización⁸. Es una distinción que importa si alguien diseña apoyos: no es el mismo problema el de quien no puede pagar la herramienta que el de quien ya la pagó y no logra que funcione.

Las dificultades con las competencias de los empleados fueron mencionadas por más de 60% de los participantes, sin importar el tamaño de la empresa, lo que sugiere que la brecha de habilidades digitales es un problema sistémico en Rumanía⁹. Y hay un área donde casi nadie reportó avances: la optimización de inventario y logística tuvo la frecuencia de mención más baja, nueve en total entre las 25 entrevistas, lo que sugiere que sigue siendo un desafío para la mayoría de las empresas pequeñas, probablemente porque requiere sistemas de gestión más complejos¹⁰.

Este estudio no nació en el vacío. La mayoría de las investigaciones sobre pymes rumanas de comercio electrónico usan encuestas cuantitativas, que dan un panorama general del nivel de digitalización pero limitan la comprensión de los mecanismos organizacionales y las experiencias gerenciales¹¹. Algunos trabajos usan enfoques cualitativos o métodos mixtos, pero son limitados y no se centran exclusivamente en el sector de comercio electrónico¹². Eso es lo que este trabajo intenta llenar: el relato de cómo se vive la adopción desde adentro.

Para dimensionar el universo del que salieron estas 25 voces: según datos de la plataforma listafirme.ro, 7,145 pymes reportaron actividad en el sector de comercio electrónico en Rumanía, de las cuales 7,021 son microempresas, 116 son pequeñas y 8 son medianas¹³. La distribución final de entrevistas fue de 13 microempresas, 7 pequeñas y 5 medianas¹⁴. Es decir, se habló con una fracción mínima de un universo abrumadoramente micro.

Sobre el terreno donde operan estas empresas: en Rumanía el comercio electrónico crece rápido, pero el nivel general de digitalización sigue por debajo del promedio de la Unión Europea —69.1% frente a 72.91%—, y persisten diferencias significativas entre regiones, sectores y tipos de empresa, impulsadas por factores como habilidades digitales insuficientes e inversiones limitadas en infraestructura tecnológica¹⁵. Ese es el suelo sobre el que se para cualquier conclusión: no es un mercado maduro digitalmente.

Así lo leemos nosotros. Hay un patrón que se repite en negocios pequeños sin personal técnico: la gente aprende a usar herramientas digitales probando y equivocándose, sin capacitación formal previa. Ese aprendizaje práctico alcanza para lo básico, pero no para lo avanzado. Si eso se sostiene, esperaríamos que dueños de tiendas en línea sin equipo técnico usen sobre todo lo gratuito o barato que ya viene con su plataforma, y que rara vez contraten especialistas o adopten sistemas complejos de análisis. ¿Cómo sabríamos que nos equivocamos? Si aparecieran muchos dueños sin personal técnico usando ya análisis predictivo avanzado o contratando consultores externos. Usted puede fijarse en cuánto de lo que sabe de sus herramientas lo aprendió solo, probando. Y si quiere dar el salto a algo más avanzado, pregunte a otros dueños cómo lo hicieron y si valió la pena antes de invertir.

Hay una segunda lectura que nos parece útil. Cuando las decisiones se toman mirando paneles y reportes automáticos, uno termina viendo el negocio a través de una pantalla diseñada para simplificar. En días normales eso ayuda; en días difíciles, la información puede ser tan incompleta que hay que adivinar. Esperaríamos entonces que los dueños sientan que entienden mejor su negocio en temporadas tranquilas, pero que enfrenten más incertidumbre cuando hay cambios bruscos de mercado o fallas técnicas. Nos equivocaríamos si los entrevistados reportaran que en crisis los paneles les dieron información clara y suficiente para decidir sin adivinar. Pregúntese qué tan seguido, en semanas malas, siente que los números que ve no le cuentan la historia completa. Tener a mano una fuente directa —clientes, empleados, proveedores— para contrastar lo que muestran las herramientas es algo que usted puede hacer esta semana.

Conviene ser claro sobre lo que este trabajo no puede decir. Es un estudio cualitativo y transversal: no permite generalización estadística, y sus hallazgos no prueban que la analítica o la inteligencia artificial hayan causado las mejoras de desempeño que los gerentes percibieron. Tampoco se comparó a quienes adoptaron con quienes no lo hicieron, ni se midió el retorno de la inversión. Los beneficios son lo que los dueños reportaron, no resultados garantizados. Y todas las empresas eran rumanas: los costos, las reglas legales y las condiciones de proveedores pueden ser distintos en América Latina, Estados Unidos o Canadá.

Otros trabajos que leímos, en resúmenes a los que tuvimos acceso, coinciden en que la inteligencia artificial se reconoce cada vez más como un activo estratégico para la innovación y la competitividad de las pymes, aunque la investigación sobre las experiencias concretas de adopción sigue siendo escasa. También leímos que la adopción de inteligencia artificial entre pequeñas empresas se aceleró desde 2020, pero que el éxito no ha sido consistente: las limitaciones de habilidades, la escasa preparación de los datos y la débil capacidad de gobierno aparecen como los frenos que más se repiten. Y leímos que el simple paso del tiempo no basta: crecer en cantidad de aplicaciones no lleva por sí solo a más eficiencia sin una implementación concreta y dirigida.

Lo que este estudio deja en sus manos es un espejo, no una promesa. Si usted maneja una tienda en línea sin personal técnico, la pregunta útil no es si la inteligencia artificial funciona en general, sino qué herramienta concreta usa hoy, quién la aprendió a manejar y qué decisión tomó la última vez gracias a ella. Empiece por ahí antes de comprar nada nuevo.

De dónde sale cada dato de contexto, y cuánto leímos de cada documento

Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The results revealed an intermediate level of digital maturity, dominated by the use of descriptive analytics, while AI applications were only partially implemented. The main barriers identified were the deficit of digital competencies and financial constraints, with variations across firm size categories."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "As these barriers are overcome, the adoption of data analytics and artificial intelligence technologies contributed to a more rigorous grounding of decision-making and to improved financial performance through sales optimisation, cost reduction, and inventory management efficiency."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Data analytics refers to the systematic use of statistical and analytical techniques to generate insights relevant to the managerial process (Mikalef et al., 2019; Chatterjee et al., 2023), while artificial intelligence denotes technologies capable of simulating cognitive processes, such as machine learning and prediction generation, contributing to the optimisation of SMEs’ financial performance (Arroyabe et al., 2024)."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Microenterprises predominantly rely on free or low-cost tools such as Google Analytics, standard reports offered by the e-commerce platform, and Excel for basic financial analyses."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Small enterprises present an intermediate level of technological endowment, integrating SaaS solutions such as CRMs, email marketing platforms, and occasionally automation solutions."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Medium SMEs demonstrate the most advanced level of adoption, implementing full ERP systems, advanced Business Intelligence platforms (Power BI, Tableau), sophisticated CRMs, and notably, predictive AI models for demand forecasting, price optimisation, and marketing personalisation."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The high costs of professional software, the complexity of technical integration between systems, the need for repeated tests and the adjustments, and lack of internal expertise represent the major identified obstacles."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The essential difference is that microenterprises are primarily constrained by financial barriers, while larger firms face challenges related to implementation and optimisation – an important consideration for support policies."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Difficulties related to employee competencies were mentioned by over 60% of the participants, regardless of firm size, suggesting that the digital skills gap is a systemic issue in Romania"
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Inventory and logistics optimisation presents the lowest coding frequency (9 mentions: 3 micro, 4 small, 2 medium), suggesting that this area remains a challenge for most small SMEs, probably due to the need for more complex inventory management systems."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Most studies on Romanian e-commerce SMEs employ a quantitative approach based on surveys (Cazabat et al., 2019; Türkeş et al., 2019; Măiţă, Pădurean and Cazacu, 2021; Iordache et al., 2023; Mihu and Herciu, 2024), providing a general overview of the level of digitalisation and its determinants, but limiting the understanding of organisational mechanisms and managerial experiences."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Some studies use qualitative approaches (González-Varona et al., 2021; Lupșa-Tătaru, Marinescu and Rădulescu, 2025) or mixed methods (Borcoși et al., 2023); however, these are limited and do not focus exclusively on the e-commerce sector."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "According to data published on the listafirme.ro platform, 7,145 SMEs reported activity in the e-commerce sector in Romania, of which 7,021 are microenterprises, 116 are small enterprises, and 8 are medium enterprises."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The final effective distribution (Table no. 1) is 25 interviews conducted until theoretical saturation was reached: 13 microenterprises (52%), 7 small enterprises (28%), and 5 medium enterprises (20%)."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - el artículo del que trata esta nota — el artículo completo — el pasaje: "In Romania, e-commerce is experiencing rapid growth, yet the overall level of digitalisation remains below the European Union average (69.1% compared to 72.91%) (EC, 2025). Significant differences also persist across regions, sectors, and types of firms (EIB, 2023; Eurofound, 2025), driven by factors such as insufficient digital skills and limited investments in technological infrastructure (Mihu and Herciu, 2024; Apostoaie et al., 2025)."

Los hallazgos de otros estudios que aquí se mencionan los conocemos por este documento, que fue el que leímos; no abrimos cada uno de esos estudios.

interviews · Amfiteatru Economic · the paper, 1 May 2026 · free

The Shop Owner With Excel and Google Analytics

In 25 interviews, Romanian online sellers said analytics and AI sharpened their decisions and their margins. The study recorded what they believed, not what their books proved.

Short version · the longer version follows, about 8 min

Ask the weeklyAI chat

Ask me about this study: who was studied, what it found, and what it does not say.

Conversations are saved for as long as weeklyAI exists, to improve the publication. Answers come in the language you write in.

The study at a glance
Who: owners and managers of Romanian online shops
How many: 25
Where: Romania
When: September and October 2025
Kind of study: interviews
Who did it: University of Craiova, Romania
The limit that matters: It recorded what owners said they noticed, not audited results or a comparison group.

Share of the 25 interviews by firm size

microenterprises52%

small enterprises28%

medium enterprises20%

How the 25 interviews were split by firm size; the study stopped at 25 rather than the 30 planned, so these are the shares of the interviews held, not of all Romanian online shops.

What held back firms of different sizes

microenterprisesagainstlarger firms

microenterprises were held back mainly by money; larger firms struggled more with putting the tools in and fine-tuning them

Next time a dashboard tells you what to do, ask what it cannot see.
weeklyAI's reading

Watch · 0:48

"We use Google Analytics and reports from the e-commerce platform. We aggregate data in Excel."

That is a microenterprise owner in Romania, and his setup is the one most of his peers described. The study behind it is small: 25 owners and managers of Romanian online shops, interviewed by telephone in September and October 2025.

The pattern followed the size of the firm. Micro businesses leaned on free tools, the platform's own reports and spreadsheets. Small firms added customer management software and Power BI. Medium firms ran full ERP systems and predictive models for demand and pricing.

Owners said decisions became faster and easier to defend. They reported steadier sales, better margins by product category, and tighter inventory control. One theme came up 42 times: how these tools changed how the business ran. Owners described faster reactions to market changes, better offers, and resources placed more efficiently.

Then came the wall. Many owners said their employees did not have the skills to use the tools well. "It is difficult to find people trained in data analysis," one small-firm manager said. Micro businesses named budget first: "Limited budget does not allow us advanced solutions."

Two things to keep straight. This study asked owners what they noticed. It did not compare their financial statements before and after, and it did not test these tools against any alternative. So it cannot show that analytics or AI caused the gains the owners described. Every firm was in Romania, where prices, rules and suppliers differ from yours.

The study also stopped at 25 interviews, short of the 30 planned, because the last three added nothing new. That is normal for this kind of research. It also means the patterns are examples to hold against your own shop, not a forecast for it.

Before you buy the next tool, look at two things: whether anyone on your team can actually read what it produces, and whether the free version you already have is answering the questions you are asking.

What this means for you

Los 25 dueños y gerentes describieron lo que notaron, no lo que muestran sus cuentas, así que no puede concluir todavía que estas herramientas causaron las mejoras. Si su equipo no tiene quien lea los reportes, o si la versión gratuita que ya usa responde sus preguntas, conviene esperar antes de pagar por algo más.

Who paid: The article says the research was conducted with the support of the Center for Interdisciplinary Research in Economics and Social Sciences, INCESA (Research Infrastructure in Applied Sciences), University of Craiova; it does not say whether the funders had any say or who lent equipment or software.

The longer versionThe passages copied from the paper, the pictures, and every source with how much of it we read · 1637 words · about 8 minRead it →Close

Romanian Shop Owners Say Analytics Sharpened Decisions and Helped Margins — But the Study Recorded Perceptions, Not Results

Twenty-five owners and managers of online stores described what changed in how they decided and what it cost them to get there.

How it could look · illustration generated by weeklyAI.watch, not a photograph

Twenty-five people who run online shops in Romania sat down for interviews about the software they use to make decisions — the reports their shopping platform produces, spreadsheet files, paid dashboard tools, customer databases, and occasionally AI tools that predict demand or suggest prices. Researchers from the University of Craiova asked them what they had adopted, why, what got in the way, and what they believed had changed in their decision-making and their financial results. The picture that came back was middling: most were using software that describes what already happened — sales by category, traffic, margins — while AI applications were only partly in place¹. The authors are Laura Vasilescu, Mirela Sichigea, Valeriu Brabete, Daniel Goagără and Dorel Berceanu, all of the University of Craiova in Romania, and the research was carried out with support from the university's interdisciplinary research centre. It appeared in Amfiteatru Economic in 2026, with interviews conducted in September and October 2025.

The obstacles the owners named were consistent: the cost of professional software, the difficulty of making systems work together, the rounds of testing and adjusting, and a shortage of internal expertise². More than 60 percent of participants mentioned trouble with employee skills, whatever the size of their firm³. Money and skills were the two recurring barriers, and they hit firms differently: the smallest firms were held back mainly by budget, while larger firms struggled more with implementation and fine-tuning⁴. The smallest shops leaned on free or cheap tools — the standard reports that come with a shopping platform, a free web analytics account, spreadsheet files for basic financial analysis⁵. Small firms sat in the middle, adding subscription tools for customer records and email marketing⁶. Medium firms went furthest, running full business-management systems, paid dashboard platforms and predictive models for demand forecasting, price setting and marketing personalisation⁷. Where those barriers eased, owners described decisions that were better grounded and financial performance improved through sales optimisation, cost reduction and more efficient inventory management⁸.

What the study did not do matters as much as what it did. It recorded what owners and managers said they perceived — not audited financial statements, not a comparison between firms that adopted these tools and firms that did not. No control group, no before-and-after measurement, no return-on-investment figure. The authors themselves note that the qualitative, cross-sectional design does not allow statistical generalisation. The sample was drawn from firms that reported the specific retail-by-internet activity code and had contact details available. Inventory and logistics optimisation drew the fewest coded mentions — nine across all 25 interviews — which the authors read as a sign that this area remains a challenge for most small firms, probably because it needs more complex systems⁹.

Here is what the technology actually is, in plain terms. Data analytics means using statistical and analytical techniques systematically to produce insights a manager can act on; artificial intelligence means software that mimics cognitive processes, such as machine learning and generating predictions, and the article frames both as contributing to the optimisation of small firms' financial performance¹⁰. The difference between the two matters for a shop owner. Descriptive analytics tells you what happened last month. A predictive model tells you what might happen next month if you price a product a certain way. The Romanian firms in this study were mostly at the first stage, with the second only partly implemented¹.

The study belongs to a family of work asking whether small firms actually get anything measurable from these tools. Most earlier studies of Romanian online sellers used surveys, which give a broad picture of how digitalised firms are and what drives that, but limit what we can learn about the mechanisms inside a business and how managers experience them¹¹. A smaller number used interviews or mixed methods, but those did not focus exclusively on the e-commerce sector¹². This study was designed to fill that specific gap. The authors also report the shape of the sector they drew from: 7,145 firms in Romania reported e-commerce activity, of which 7,021 were microenterprises, 116 small and 8 medium¹³. The final sample of 25 interviews — 13 micro, 7 small, 5 medium — was stopped when the last three consecutive interviews produced no new codes or themes¹⁴.

Other work we read adds useful texture. According to the summary of a study of fourteen women leading established small and medium businesses across various industries in Chile — we could read only the summary, the full paper is behind a subscription — participants used AI mainly to support customer relationship management and described limited readiness as a constraint tied to learning needs¹⁵. We also read, again only in summary form, a practitioner review of small-firm AI adoption published between 2020 and 2025, which found that small firms most often use AI for customer engagement, marketing content and administrative automation, and report productivity gains when tools are paired with clear workflows and human oversight¹⁶. The same review names limited skills and weak data readiness as the most consistently reported barriers¹⁷. And the summary of a survey of 228 small and medium firms reports that different applications of AI had a statistically significant positive impact on labour productivity and on reducing operating costs, while the impact on business process efficiency was moderate and partially limited¹⁸¹⁹; that study also found that operational applications such as automation and data analysis mattered more than AI used in managerial decision-making²⁰²¹. The Romanian interviews sit alongside these findings rather than confirming or contradicting them — different countries, different methods, different questions.

Here is how we read it. The owners in this study described a change in how decisions felt: less speculative, faster, more grounded in numbers. That is a real change in a business, and it is also a change that can be mistaken for a change in results. A dashboard that makes you feel certain is not the same as a dashboard that is right. The pattern we would expect in shops like yours is this: without technical staff, you will treat the tool's output as an authority — the platform's built-in report, the AI tool's suggestion, the polished chart — and act on it without asking whether the underlying data is complete or whether the metric is the one that matters. The study's finding that confidence rose with data use may partly reflect that deference rather than genuine improvement. We would know we are wrong if owners routinely asked what a number does not capture, tested a tool's recommendation against their own experience, and kept a short record of decisions they made against the tool's advice. That habit costs nothing and is the cheapest protection you have: before acting on a dashboard or an AI recommendation, ask one question out loud — what would have to be true for this number to be wrong?

There is a second pattern worth watching, and it concerns what you give up when you buy your way in. Small online shops with no technical staff tend not to commit to a single analytics or AI provider; they keep free platform reports alongside a paid tool, run parallel spreadsheets, and stay on month-to-month arrangements. The cost does not appear as a line item. It appears as duplicated work and slower decisions, and it is worst for the smallest firms. We would know we are wrong if micro and small shops were standardising on one integrated tool, dropping their spreadsheets, and signing longer contracts without keeping a fallback. Count how many separate places you check before making a pricing or inventory decision. If it is more than two, that is a real cost — not just an inconvenience — and it is worth pricing out one integrated option against the hours you spend reconciling them. And if you hand analytics setup, ad management or AI tool configuration to an outside agency or freelancer, watch where the decision-making actually sits. Ask them to walk you through one recent recommendation in plain language, including what data it came from and what could make it wrong. If they cannot or will not, treat that as a warning. Also check what data they hold and what happens to it if you leave.

The study's own recommendations for Romanian firms were to invest in analytics infrastructure such as integrated systems and automation platforms, to develop employees' digital skills including certifications in data management and AI models, and to optimise logistics, inventory and financial operations. The authors point to public support programmes, professional training and competency ecosystems as ways to accelerate adoption across small and medium firms, and note that microenterprises relying on free or limited-functionality tools may see the performance gap widen. Future research they call for includes comparative quantitative studies across sectors, longitudinal work on AI adoption, and evaluations of the return on digital investments — precisely the measurements this study could not provide.

What this makes possible for you is narrower and more useful than a verdict on whether analytics works. You now have a set of questions to bring to your own business: which of your tools merely describes the past, and which one actually predicts something you could act on? Who would notice if the numbers were wrong? What would it cost you to consolidate the places you check before a pricing decision? And what would you need to see, in your own numbers, before you believed a tool had paid for itself? The Romanian owners could not tell you that — they could only tell you how it felt. Next time a dashboard tells you what to do, ask what it cannot see.

Where each piece of context comes from, and how much of it we read

Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - the article this story is about — the whole article — the passage: "The results revealed an intermediate level of digital maturity, dominated by the use of descriptive analytics, while AI applications were only partially implemented. The main barriers identified were the deficit of digital competencies and financial constraints, with variations across firm size categories."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - the article this story is about — the whole article — the passage: "The high costs of professional software, the complexity of technical integration between systems, the need for repeated tests and the adjustments, and lack of internal expertise represent the major identified obstacles."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - the article this story is about — the whole article — the passage: "Difficulties related to employee competencies were mentioned by over 60% of the participants, regardless of firm size, suggesting that the digital skills gap is a systemic issue in Romania"
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - the article this story is about — the whole article — the passage: "The essential difference is that microenterprises are primarily constrained by financial barriers, while larger firms face challenges related to implementation and optimisation – an important consideration for support policies."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - the article this story is about — the whole article — the passage: "Microenterprises predominantly rely on free or low-cost tools such as Google Analytics, standard reports offered by the e-commerce platform, and Excel for basic financial analyses."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - the article this story is about — the whole article — the passage: "Small enterprises present an intermediate level of technological endowment, integrating SaaS solutions such as CRMs, email marketing platforms, and occasionally automation solutions."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - the article this story is about — the whole article — the passage: "Medium SMEs demonstrate the most advanced level of adoption, implementing full ERP systems, advanced Business Intelligence platforms (Power BI, Tableau), sophisticated CRMs, and notably, predictive AI models for demand forecasting, price optimisation, and marketing personalisation."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - the article this story is about — the whole article — the passage: "As these barriers are overcome, the adoption of data analytics and artificial intelligence technologies contributed to a more rigorous grounding of decision-making and to improved financial performance through sales optimisation, cost reduction, and inventory management efficiency."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - the article this story is about — the whole article — the passage: "Inventory and logistics optimisation presents the lowest coding frequency (9 mentions: 3 micro, 4 small, 2 medium), suggesting that this area remains a challenge for most small SMEs, probably due to the need for more complex inventory management systems."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - the article this story is about — the whole article — the passage: "Data analytics refers to the systematic use of statistical and analytical techniques to generate insights relevant to the managerial process (Mikalef et al., 2019; Chatterjee et al., 2023), while artificial intelligence denotes technologies capable of simulating cognitive processes, such as machine learning and prediction generation, contributing to the optimisation of SMEs’ financial performance (Arroyabe et al., 2024)."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - the article this story is about — the whole article — the passage: "Most studies on Romanian e-commerce SMEs employ a quantitative approach based on surveys (Cazabat et al., 2019; Türkeş et al., 2019; Măiţă, Pădurean and Cazacu, 2021; Iordache et al., 2023; Mihu and Herciu, 2024), providing a general overview of the level of digitalisation and its determinants, but limiting the understanding of organisational mechanisms and managerial experiences."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - the article this story is about — the whole article — the passage: "Some studies use qualitative approaches (González-Varona et al., 2021; Lupșa-Tătaru, Marinescu and Rădulescu, 2025) or mixed methods (Borcoși et al., 2023); however, these are limited and do not focus exclusively on the e-commerce sector."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - the article this story is about — the whole article — the passage: "According to data published on the listafirme.ro platform, 7,145 SMEs reported activity in the e-commerce sector in Romania, of which 7,021 are microenterprises, 116 are small enterprises, and 8 are medium enterprises."
Vasilescu, L., Sichigea, M., Brabete, V. et al. (2026). Decision-Making and Financial Performance in E-Commerce SMEs: Perceptions on Data Analytics and Artificial Intelligence. Amfiteatru Economic. https://doi.org/10.24818/EA/2026/72/650 - the article this story is about — the whole article — the passage: "The final effective distribution (Table no. 1) is 25 interviews conducted until theoretical saturation was reached: 13 microenterprises (52%), 7 small enterprises (28%), and 5 medium enterprises (20%)."
Ireta-Sánchez JM, Berkatsashvili M. (2026). Artificial Intelligence Adoption in Established Women-Led SMEs: A Multiple-Case Study of Competencies, Barriers, and Business Process Applications. Systems. https://doi.org/10.3390/systems14091120 — only the abstract - the full text could not be fetched — the passage: "Participants reported utilising AI primarily to support customer relationship management (CRM), while also describing limited AI readiness as a constraint associated with learning needs and experiences of AI adoption."
Minifie J, Arango Mesa S, Ayer S. (2026). Incorporating Artificial Intelligence in Small Businesses: A Practitioner-Oriented Literature Review (2020–2025). Small Business Institute Journal. https://doi.org/10.53703/001c.168695 — only the abstract - the full text could not be fetched — the passage: "Small firms most often use AI for customer engagement, marketing content creation, and administrative automation, and they report productivity gains when tools are paired with clear workflows and human oversight."
Minifie J, Arango Mesa S, Ayer S. (2026). Incorporating Artificial Intelligence in Small Businesses: A Practitioner-Oriented Literature Review (2020–2025). Small Business Institute Journal. https://doi.org/10.53703/001c.168695 — only the abstract - the full text could not be fetched — the passage: "Limited skills and weak data readiness are the barriers reported most consistently."
Bolfek M, Rajko M, Bolfek B. (2026). The Economic Effects of Artificial Intelligence Adoption in Small and Medium-Sized Enterprises. World. 10.3390/world7060103 — only the abstract - the full text could not be fetched — the passage: "The research results show that different applications of artificial intelligence have a statistically significant, positive impact on labor productivity and on reducing operating costs."
Bolfek M, Rajko M, Bolfek B. (2026). The Economic Effects of Artificial Intelligence Adoption in Small and Medium-Sized Enterprises. World. 10.3390/world7060103 — only the abstract - the full text could not be fetched — the passage: "In contrast, their impact on business process efficiency is moderate and partially limited."
Bolfek M, Rajko M, Bolfek B. (2026). The Economic Effects of Artificial Intelligence Adoption in Small and Medium-Sized Enterprises. World. 10.3390/world7060103 — only the abstract - the full text could not be fetched — the passage: "The operational application of artificial intelligence, such as automation and data analysis, has proven to be the most important factor in economic effects."
Bolfek M, Rajko M, Bolfek B. (2026). The Economic Effects of Artificial Intelligence Adoption in Small and Medium-Sized Enterprises. World. 10.3390/world7060103 — only the abstract - the full text could not be fetched — the passage: "At the same time, its application in managerial decision-making also has a significant, but somewhat weaker impact."