Ekşi Sözlük is a Turkish collaborative dictionary and forum where users write entries under topics, including topics about brands. What is written there about your chain looks nothing like a review form or a complaint file: topics stay open for years, entries can be sarcastic or crude, and store experiences, campaign criticism and unrelated chatter sit side by side on the same page. This post explains how Ekşi Sözlük brand monitoring is set up in BranchVoice, how entries are collected and separated from noise, and how store mentions are linked to the right branch.
What gets said about a brand on Ekşi Sözlük
For chains, Ekşi Sözlük is different from other sources in two ways. First, content accumulates under brand topics: next to the "brand name" topic you will find sub-topics such as "brand name return policy" or "brand name mall store", and unless all of them are tracked, part of the picture is lost. Second, the authors are often observers rather than customers: an entry may describe a real store visit, or it may just be mocking an advert.
That is why the real work in monitoring Ekşi Sözlük is not collecting but filtering. What an operations team needs is to find the sentence "I waited twenty minutes at the till in the Izmir store" among hundreds of entries and get it to the responsible regional manager.
Finding the right topics
In the store import wizard you pick Ekşi Sözlük as the source. Three kinds of input are accepted:
- Keywords: the brand name and its variations. The wizard searches Ekşi Sözlük for the keywords and lists the matching topics; you choose which ones to follow. This is how you discover sub-topics you did not know existed.
- Topics: topic names you already know can be added directly.
- Start URLs: you can paste the link to a specific topic.
You then set the maximum number of entries and the page limit per topic. There is no candidate review step; the first fetch starts as soon as the source is created. The overall import logic is covered in importing your store list.
How entries are collected
The first fetch brings in the history of the selected topics up to the limit you set. Every entry is stored in BranchVoice as a review, with the topic name added as a tag, so in the review list you can see all entries from the "return policy" topic with a single filter. Since Ekşi Sözlük has no star ratings, sentiment is derived from the text alone.
Scheduled sync is deliberately kept small for Ekşi Sözlük: each run checks only the newest page of a topic and brings in a limited number of new entries. This prevents a topic that has been open for years from being re-scanned every week; data-provider cost stays small and only new content reaches the dashboard. You follow the runs on the sync jobs page and trigger one manually when needed; see scheduled review sync for details.
Filtering out the noise
Not every entry collected from Ekşi Sözlük is customer feedback. This is where BranchVoice's AI analysis applies a first screen: entries with no connection to the brand or to the customer experience are marked as irrelevant and left out of the analysis, so they never enter dashboard statistics. A joke about a TV advert or general chatter under the topic is removed this way.
Relevant entries get additional tags. Text containing slang or profanity is flagged separately, so the summaries that reach a regional manager highlight the substance rather than the wording. Content that looks suspicious, for instance organised or artificial, is tagged as well, and anomaly detection separately reports suspicious entries piling up in a topic within a short time. The full set of classification questions is described in AI review classification.
Store mentions and action
Like Şikayetvar complaints, Ekşi Sözlük entries are first attached at brand level to the company's virtual "General" store. If AI finds a branch clue in the text, such as a mall name, a district or a store name, it compares the entry against candidate stores from your list and moves it to that store when the decision is confident enough. Entries that cannot be resolved stay in the General pool; high-risk ones are placed in the manual link queue, where a team member picks the store by hand.
This distinction matters operationally. An entry linked to a store appears in the regional manager's view in the same list as reviews from other platforms and can trigger alert rules. Entries in the General pool are head-office matters: campaigns, pricing policy, brand perception. The customer experience team follows those through the topic breakdown.
A typical example for a fashion chain: entries piling up in the "return policy" topic stay in the General pool and feed the policy discussion, while the entry "spent hours trying to return something at the mall store in Ankara" from the same topic is moved to that store and lands in its weekly summary.
Next steps
Ekşi Sözlük may not match Google's volume, but it is the source that shows brand perception in its most candid form, and because AI filters out irrelevant content, nobody on your team has to read the topics by hand. To see your brand's topics broken down by store in BranchVoice, request a demo.