Adjust has published Mobile app trends spotlight edition: Türkiye 2026, a benchmark report covering gaming, finance and shopping apps from January 2024 to July 2026, with MENAT and global comparisons throughout. It is one of the clearest pictures available of how people in Türkiye actually use apps: how often they open them, how long they stay, and how quickly they leave.
We read it the way we read any dataset before a client engagement: what does it say, what does it change about where to look for growth, and where does it need careful handling. This article is our commentary. All figures below come from the Adjust report unless we say we calculated them ourselves; third-party figures are the ones Adjust cites.
The market in five numbers
- Reach is close to total. Of 87.7 million people, 81.9 million have active mobile connections (93.3%) and 77.5 million use the internet regularly (88.3%). Median age is 33.5.
- E-commerce is large and still growing fast. The report cites Ministry of Trade ETBİS data putting 2025 e-commerce volume at $115.43 billion, up 52.2% year on year.
- Türkiye makes games more than it buys them. Turkish developers earned $2.76 billion worldwide in 2025, against $347 million spent by players in Türkiye.
- Session growth in H1 2026 (YoY): gaming +6% (global +6%), e-commerce +4% (global +0.3%), finance +2% (global +29%).
- Retention versus benchmarks: finance beats MENAT and global at every point to day 30; gaming trails global at every point.
Those headlines are useful, but the more actionable material sits in the segment tables. Six things stood out to us.
1. In gaming, install share and engagement share are two different markets
The report breaks down H1 2026 gaming installs and sessions by product model and by genre. Dividing each segment's share of sessions by its share of installs gives a simple engagement multiplier. The ratios below are our calculation from Adjust's shares.
| Segment | Share of installs | Share of sessions | Sessions ÷ installs |
|---|---|---|---|
| Mid-core (product model) | 12% | 51% | 4.25× |
| Action (genre) | 10% | 33% | 3.3× |
| Sport games (genre) | 7% | 15% | 2.1× |
| Strategy (genre) | 6% | 11% | 1.8× |
| Casual (product model) | 17% | 19% | 1.1× |
| Puzzle (genre) | 17% | 13% | 0.8× |
| Hybridcasual (product model) | 18% | 12% | 0.7× |
| Hypercasual (product model) | 53% | 18% | 0.3× |
Our read: more than half of all gaming installs in Türkiye go to hypercasual titles that produce under a fifth of sessions, while mid-core turns an eighth of installs into half of all play. A team that ranks campaigns, creatives or channels by cost per install is optimising for the bottom of this table. Sessions per install at day 7, by acquisition source, is a far better steering metric, and it is cheap to build from any attribution tool.
One more point on this data. Average gaming session length in Türkiye rose to 33.69 minutes in H1 2026, from 31.11 in 2024. Part of that rise is mix: mid-core sessions average 43.32 minutes and carry 51% of all sessions, so any shift of play towards mid-core lifts the average without any single game getting stickier. Casual is the segment where the per-session change is real, from 23.18 to 26.24 minutes in one year.
2. Turkish players leave faster, then play longer
| Gaming retention, H1 2026 | Türkiye | Global | Türkiye as % of global (our calculation) |
|---|---|---|---|
| Day 1 | 20% | 27% | 74% |
| Day 7 | 6% | 13% | 46% |
| Day 30 | 1% | 5% | 20% |
The gap does not stay constant; it widens with every checkpoint. Yet Turkish sessions are the longest of the three regions (33.69 minutes against a global 30.52). MENAT shows the same shape, with day 1 at 21%.
Our read: this is the profile of a small, committed core surrounded by a large volume of users who sample and leave. The largest relative loss happens between day 1 and day 7, which is where we would concentrate product experiments: the second-session trigger, the first progression unlock, and notification timing in the first week. Content depth is clearly not the problem for the players who stay.
3. Finance: low session growth, the best relative retention, and short sessions that are a feature
Finance sessions in Türkiye grew only 2% YoY in H1 2026, far behind MENAT (12%) and global (29%). But retention is the report's standout: day 1 at 14% against 13% (MENAT) and 12% (global), and day 30 at 4%, double the global 2%. Sessions are short, at 6.12 minutes against a global 8.27.
Our read: short sessions plus high retention is what a well-functioning tool looks like. People open a banking app, pay, check, transfer, and leave, then come back. Finance teams should not borrow engagement targets from content or gaming apps; the metric to improve is time to complete a task, and the funnel to test is the one between opening the app and finishing the most common three tasks.
The subvertical table adds a second signal:
- Bank: installs +31%, sessions +8%
- Insurance: installs +24%, sessions +14%
- Stock trading: installs +41%, sessions +43%
In banking, installs are growing almost four times faster than usage. That points to an activation question: are new installs coming from campaigns, account-opening incentives or switching drives that do not become regular use? The practical test is to cohort bank installs by source and measure the share that complete a first real transaction within seven days. Stock trading is the opposite case, where usage keeps pace with acquisition.
There is also a piece of arithmetic worth noticing: all three listed subverticals grew sessions by more than the 2% finance total. For the total to be 2%, some of the subverticals not listed must have grown by less than 2% or shrunk. The report does not say which.
4. E-commerce growth is coming from people who already have the app
E-commerce sessions in Türkiye grew 4% YoY in H1 2026, against 0.3% globally, and the report describes it as a new high. Installs did not follow the same path: the monthly chart shows installs running below sessions through 2026, with May and July the best months at +7% above the H1 average. Day 1 retention was 14% (MENAT and global 13%), converging with both benchmarks at 6% by day 7. Sessions average 8.63 minutes, below the global 10.32.
Our read: session growth without matching install growth means the existing base is opening apps more often. For Turkish e-commerce teams, the conversion levers are mostly inside the app: home-screen personalisation, search, recently viewed and saved items, reorder flows and CRM that brings people back on the right day. Shorter-than-global sessions suggest frequent, quick visits, so the design goal is to make the second and third visit of the week productive, not to make each visit longer.
5. Plan experiments around the calendar the report shows
Each vertical has a clear seasonal shape:
- Gaming: January and March peak every year. In 2026, January installs were 19% above the H1 average and sessions 11% above; March was +11% on both.
- Finance: January, March and July sessions ran 3–4% above average, with January installs +10%.
- E-commerce: the late-year run dominates. November 2025 sessions were 15% above the annual average, December +10%, October +6%.
Our read: peak months bring a different traffic mix, with more deal-seekers, more first-time buyers and more promotional intent. A test won in November often does not hold in February. For e-commerce teams reading this at the end of September, the practical plan is to finish Q4 experiments by the end of October, ship the winners before November, and use the peak to confirm results rather than to discover them. Gaming teams have the equivalent window in November and December, ahead of the January peak.
6. AI in Türkiye: visibility largely means visibility in ChatGPT
Citing Sensor Tower, the report shows generative AI app downloads in Türkiye rising from 33.3 million to 82.9 million in 2025, with time spent up 375% against 259% globally. More striking for marketers: ChatGPT receives 94.5% of AI-related web traffic in Türkiye, against 81% worldwide. AI is also moving into apps people already use: Garanti BBVA's LLM-based assistant Ugi handled 73 million conversations for 8 million customers, and Trendyol runs AI behind its seller assistant and customer search.
Our read: for Generative Engine Optimisation in Türkiye, one assistant carries most of the referral surface today. A Turkish-language prompt panel run against ChatGPT covers most of what matters for brand and product visibility right now. That concentration is a snapshot, not a law, so the panel should be built to add Gemini and others without redesign.
How to read the report carefully
The report is strong, and like any benchmark it has limits. These are the ones we would keep in mind before quoting it in a planning deck:
- It is Adjust's panel, not the whole market. The methodology says so directly: data comes from apps Adjust measures, drawing on its top 5,000 apps and its full dataset.
- Growth figures are indexes, not volumes. The report gives percentages and shares, not absolute installs or sessions, so it can tell you direction and relative size, not market size.
- The gaming segment tables do not reconcile with the headline. Applying the product-model session growth rates (casual +76%, hybrid casual +13%, hypercasual +1%, mid-core +3%) to their H1 2026 session shares implies total growth of roughly 13%, not the 6% headline (our calculation). The most likely explanation is that segment tables and headline come from different samples, which the two-dataset methodology allows. Use segments for direction, not to size an opportunity.
- Session-length periods are not like for like. 2024 and 2025 are full years; H1 2026 is half a year. In e-commerce especially, where Q4 is the peak, comparing a half year with full years can move averages on its own.
- "Strongest retention" is relative. Finance beats its benchmarks by the widest margin, but in absolute terms gaming's day 1 retention (20%) is higher than finance (14%) or e-commerce (14%).
- Retention is not revenue. None of these curves tell you what a retained user is worth, which is what budget decisions ultimately need.
What we would do with this report this quarter
- Gaming: replace CPI as the primary campaign metric with day 7 sessions per install by source, and put the next three product experiments between day 1 and day 7.
- Finance: measure time to complete the top three tasks, and cohort new bank installs by source to find which ones reach a first transaction within a week.
- E-commerce: lock the Q4 test calendar now, finish tests by the end of October, and shift effort from acquisition to return-visit experience: search, saved items, reorder and CRM timing.
- All verticals: build your own benchmark from your data against these numbers, and set up a Turkish-language ChatGPT visibility check for your brand and category.
Frequently asked questions
What does the Adjust Türkiye 2026 report cover?
Installs, sessions, session lengths and retention for gaming, finance and shopping/e-commerce apps in Türkiye, compared with MENAT and global benchmarks, using Adjust data from January 2024 to July 2026. It also includes a section on generative AI apps drawing on third-party sources.
Is app retention in Türkiye higher or lower than global?
It depends on the vertical. Finance apps in Türkiye retain better than global at every point to day 30, e-commerce is slightly ahead on day 1 and level by day 7, and gaming trails global throughout, with the gap widening from day 1 to day 30.
Which Turkish app vertical grew fastest in H1 2026?
Relative to global benchmarks, e-commerce: sessions grew 4% YoY against 0.3% globally. Gaming matched the global rate at 6%, and finance grew 2% against 29% globally.
If you want to see how your own app or mobile site compares with these benchmarks, and which funnel step is costing you the most, Switas runs growth and CRO programmes for e-commerce, finance and consumer apps in Türkiye and beyond. The full report is available from Adjust.
